Anna (00:00:07):
Welcome to zero knowledge, a podcast, where we talk about the latest in zero knowledge research and the decentralized web.

Anna (00:00:16):
The show is hosted by me, Anna,

Fredrik (00:00:19):
And me Fredrik.

Anna (00:00:27):
In this week's episode, guest host Tarun and I chat with Alex Svanevik, CEO and co founder of the company Nansen. We cover crypto analytics, DeFi analytics, as well as discuss the push and pull between a need for privacy and the need for blockchain, transparency, and data. Now, before we start in, I want to say thank you to this week sponsor Aleo. Aleo is the first platform for fully private applications. They use blockchain and zero knowledge cryptography to deliver a web experience that is both personal and private. With Aleo developers can write private applications without a background in blockchains or any expertise in cryptography. The company consists of a world class team of cryptographers engineers and designers, and they've recently released the developer preview one, an early peek at what the future of the web could look like. The release introduces a new programming language called Leo as well as a new community driven package manager for Leo and a new development environment or IDE called Aleo Studio. We recently had Howard Wu, the co founder of Aleo on the show for an interview. If you haven't yet checked it out and want to find out more, I highly recommend giving it a listen. I'll provide a link in the show notes as well as a link to the Developer Preview that you can learn more about Aleo and start contributing today. So thank you again, Aleo. Now here is our interview with Alex from Nansen.

Anna (00:01:54):
So this week I'm happy to invite Tarun back to the show as a guest host. Hi Tarun.

Tarun (00:01:59):
Hey, how are you?

Anna (00:02:00):
Good. Nice to see you back here.

Tarun (00:02:03):
It's always good to be on the podcast.

Anna (00:02:06):
In our conversation today. We're going to be touching on crypto analytics, but we're specifically going to be looking at DeFi analytics and we have a really good guest to do that with. We have with us Alex Svanevik who's the CEO and co-founder of Nansen welcome, Alex.

Alex (00:02:21):
Thank you. Nice to be here.

Anna (00:02:23):
Before we started on you cause obviously we want to learn a little bit about your background, the name Nansen. What is that name? What does that mean?

Alex (00:02:31):
That's a really good question. Actually, we took it from an Explorer who's called Fridtjof Nansen. He's an original Explorer, Polar Explorer, and he sort of embodies the values that we have in the company of curiosity and courage. And he had many famous expeditions. He led an expedition called Fram, which made it the furthest North that anyone has ever been. Cool. So you did lot of other interesting things and he actually helped develop the theory of the Neuron. Oh wow. So see, he was also a scientist, so, so he's the inspiration for the name. It's also pretty easy to pronounce in all languages, which is good.

Anna (00:03:11):
So what got you interested in this particular space? Tell us a little bit about the path that led you to founding Nansen.

Alex (00:03:20):
Yeah, so my own personal background is in machine learning and data science. I've worked about 10 years in startups larger corporations. And I also did a few years in management consulting and then in 2017 is when I transitioned into working full time in crypto. So I discovered that there was an abundance of data in crypto and an openness around data, but I found that the space was still quite immature when it came to analytics products. And so for me personally, that meant that I could create some value in the space.

Anna (00:03:54):
Was that the first project you worked on? Like did you immediately, start in the analytics side of things?

Alex (00:03:59):
Yes. I mean, I think, you know, like a lot of others, I started first looking at, crypto investments, you know, this was right before the ICO boom and bust. And so I was actually looking quite a bit at ICOs and like what could drive success rates and multiples in terms of valuations and so on. So I did a small project around ICOs, so a couple of other people, but very early, I started looking at on-chain data, which I found to be the most interesting aspect of crypto.

Anna (00:04:29):
Cool.

Tarun (00:04:30):
You know, one thing I think that's kind of interesting is a lot of the episodes on this podcast focus on sort of future facing changes to blockchains like new blockchains, blockchains have new privacy tools or are more scalable, but you know, you've really focused on kind of retrospective analysis of blockchains and then doing things in that world. How long have people been really doing analytics within crypto and kind of, yeah, what's actually the history. Cause I feel like, you know, people have heard of chain analysis or they've heard of kind of some of these investigations, but I don't think people actually totally know how this kind of space got started.

Alex (00:05:11):
So definitely my focus has been on the Ethereum ecosystem. And so there's obviously a history before Ethereum when it comes to onchain data and you mentioned analysis, which has been, I think you could, you could say controversial player in the crypto space, they've focused a lot on having customers outside of the crypto space, right? Like law enforcement regulatory bodies, government agencies, and so on, I think with the Ethereum there's been maybe a couple of other analytics companies that are focused more on the actual crypto entities as the customer base. And so that's definitely the case with Nansen. So I would say that from, you know, the, the ICO boom and bust, there were a lot of analytics companies that of arose in that era. And then many of them transitioned into more like regulatory focus. So you might've been looking at kind of tokens and so on in 2017, but then as the interest for these tokens kind of died in 2018 and 19, many of them pivoted towards more B2B analytics and even, you know, regulatory applications like compliance. So I think there are some, you know, some examples of companies doing that. And I think, you know, if you look at the whole analytics space, there's probably been a bias towards Bitcoin, probably partially because of that, because you know, most of the customers buying analytics and data were actually, you know, on the B2B side focusing on Bitcoin,

Anna (00:06:45):
I mean, Bitcoin analytics in the case of Bitcoin, I guess all that you'd be analyzing is just like, I mean, it's purely account names and transaction information, I suppose, whereas with Ethereum, because there's like Dapps working on top of it, like when you do analytics on that, are most of those analytics still just on like value transactions or is it just on like who the accounts belong to or do you see already like a more sophisticated analytics idea happening there?

Alex (00:07:15):
Yeah, that's a great question. In our case and actually with all of onchain analytics that you do with Ethereum, you can look at much more than just value transfer. You can look at interactions with smart contracts. You can look at, for example, ownership of, you know, multi-sigs membership in different DAOs, you know, participation rates and governance. You know, it's a very rich ecosystem in that sense. And the cool thing about Ethereum is that it's actually not that hard to do onchain analytics. For example, you have events that get emitted from smart contracts and these, you can actually pull out from just reading out the logs from the blockchain. So those are some things we look at the unique thing about Nansen in particular is that we also have this off chain contextual layer on top of the product. So we don't only have kind of vanilla onchain data. We also enrich the on chain data with contextual information about, for example, which exchanges are involved in the transactions, which DeFi protocols are there. So if you want it to learn, you know, how does money flow around the different real world entities Nansen is probably the best place to go to find out.

Anna (00:08:25):
Before we dive more into Nansen itself. I'm curious to hear, like, what are the other projects I'm wondering if there's projects that I may have heard of that are doing that on Ethereum looking deeper than just transaction data, just to give some kind of maybe some reference points to people.

Alex (00:08:41):
Yeah, absolutely. So some other companies in the space are Bloxy for example, which I think focused mostly on API access, they also have a great website with lots of different, you know, insights that you can pull out of. There, you have Glassnode, which is another one doing analytics, which a lot of people use to write SQL queries on, Elementus is another one, which is a bit more on the B2B side. So there's, there's quite a few DeFi analytics companies in the space and you know, many of them have an on chain focus and some also focus more on the off chain market data, which would be the oldest like CoinMarkeCap, Coingecko, and so on, which are more on the, on the market side of things and off chain, centralized exchanges, et cetera.

Tarun (00:09:26):
So I think, you know, one interesting role that we've seen is that kind of blockchain analytics has, has historically, as you've pointed out, in Bitcoin, really been focused on analyzing crime and sort of trying to identify individual users at the cost of reducing fungibility and making privacy more difficult. But I think in DeFi there's because of kind of this richer data that you've kind of intimated, there's a bigger role for analytics, especially in terms of understanding smart contracts, state and how certain external entities are interacting with blockchain. So where do you kind of see the role of, of analytics within defy and how that's going to grow?

Alex (00:10:12):
I think there's at a high level, when we think about, you know, the user segments of Nansen, for example, which I think sort of maps out quite nicely to the entire crypto ecosystem, that's really two main purposes. Either you use analytics as an advantage when you do investing in trading, which is a big use case for it on its own crypto is interesting because it has very low transaction costs as a market, but it's also very inefficient still, which means that information is incredibly valuable. If you spot, you know an inefficiency, in many cases, you can profit from it within seconds. So that that's definitely one use case. The other use case is for the kind of crypto teams and DeFi project teams themselves to make their products better. You know, this could be trying to better understand the different wallets that are in interacting with our smart contracts.

Alex (00:11:07):
What other contracts are of those using, even if you don't know, you know, the identity, you don't need to know who they are, but the fact that you have a persistent record of that account can help you in a better understanding, you know, what other protocols they use and so on and so forth. So just where you might see, you know, analytics on web users in a web 2.0 world, you can see analytics on smart contract usage on wallets, even if you don't know who they are, but you have a persistent record over time that, you know, maps to some kind of entity in the real world. So I think those are the main use cases for investing and trading and really to make products better in the space. And, you know, from my own experience, speaking with customers at Nansen, but also just seeing usage of other products that we've, that we've talked about. I think those are the main use cases that people use analytics products for. You could also say there's like an entertainment value. So some people just like, you know, lurking and like, you know, checking out different wallets or different projects and seeing, you know, what's actually happening. But very often it borders onto the kind of investing use case for retail investors.

Anna (00:12:20):
I think you've also been cited by journalists as sometimes like if you're doing a report or something like that, they maybe would use this.

Alex (00:12:26):
Yeah. That's, that's a great point. That's actually a third use case. Yeah. That's normally we talk about three different customer segments and that's, that's actually the third one, which is still pretty small, to be honest in terms of numbers, but it is really important like strategically. And, and when you think about the impact that some of these researchers can have, definitely.

Anna (00:12:46):
Hm. What's the big difference here? You know, when you're doing DeFi and you're doing analytics on decentralized financial instruments or, you know, contracts, what have you, is there any version of this in traditional finance that you are looking at and kind of trying to replicate, or is it completely crypto native? Are you actually like, and this is also future and since you come from that world, but like, are there a lot of analytics tools like this in traditional finance?

Alex (00:13:14):
I'd actually love to hear Tarun´s take on this first before I go, actually.

Tarun (00:13:17):
Yeah. I mean, I think the data transparency doesn't exist as much. So like most of your tools that you're using are like trying to backwards infer what, you know, people's behavior or like their identity and based on, you know, how often they send certain sized orders or something to an exchange, you're like, "Oh, okay, this is a Whale who always rounds all of their order sizes to 27 or multiples of 27 or something" Right. Versus I think certainly in Ethereum, given that it's only pseudonymous and people tend to not make new addresses very often, you tend to be able to do a lot more rich analytics. And I think what we've seen in, in the crypto world is that's changed the market behavior quite a bit. When everyone is able to observe you.

Tarun (00:14:09):
I think there was a re a controversy last week where a large exchange owner who shall not be named was publicly shorting or like a trading firm associated with him was publicly shorting a coin. And they basically got short via DeFi protocol and they had an angry mob go after them on Twitter, which in normal finance, you, you know, the lender would have to leak your name to the press, which would then, you know what I mean? It's not like a, at least in my view, this is like a little richer cause the market structure is impacted by the analytics versus the, in normal finance it's the analytics exists in spite of the market structure.

Anna (00:14:48):
You're saying that behavior is probably altered because you are public, not just the behavior of the reaction, but like the behavior of the actor in the first place. If they have the foresight to realize that like they will be seen maybe slightly differently.

Tarun (00:15:02):
Yeah. I think Alex probably knows way more examples than me, of people using their product to, to kind of build a narrative like that.

Alex (00:15:10):
You know, we have customers who've spotted, you know, other entities accumulating a certain asset and then basically they've accelerated their own purchasing of said asset, you know, which in, in their own words, save them execution costs of maybe $250,000. Right. So, so just as an example, but I don't think we've seen that much of the kind of bluffing and so on that I think we will see, it's not clear that, you know, the entity that Tarun mentioned, it's not clear to what extent they would actually, you know, let's say sell the full amount of tokens that are being sent to some centralized exchange. And so you could imagine scenarios where an entity would do that in order to spoof, almost do an onchain spoof, right.

Alex (00:15:55):
Where basically the market might get frightened and start selling off and they get a better price than buying it up somewhere else. So it becomes kind of this poker game, almost of misleading people onchain which, which is interesting. And it's like its own kind of new dimension of game theory. That's interesting. But, but most of that, I think people love the, they have a fascination with, with onchain, I think, cause it's, it feels very futuristic and mysterious in a way that you can just see this like massive transactions and they very often go viral, right? Like if you see someone make a, a one Bitcoin transfer of like, you know, billions of dollars, so whatever it is, they tend to go viral. So I think there's kind of like an innate kind of curiosity or fascination among people as well, looking at onchain data and all these things come together in this wacky thing called crypto, which is like part entertainment, part finance.

Anna (00:16:48):
That's so funny. It's true. It's like you hear sometimes about the, these big movements and then there's all these, the speculation and narratives and storytelling around like what they think that might be. And sometimes you'll have sort of an expert weigh in and just be like, Oh, it's just this, don't worry about it. But then like, there's obviously the conspiracy theorists are like, Oh, but who knows? But what, from your position, like what you see, obviously you've built like an analytics tool, but do you also, I'm assuming you're paying attention to this space and the dashboards as well. Like what, what are you looking at? What would you look at? What do you think is like the most interesting thing for people to pay attention to?

Alex (00:17:28):
Yep. So I'm definitely the biggest user of Nansen in terms of hours. Like I'm a dog through the product literally every day. So again, it comes back to those use cases, right? For the three segments. Like if you're an analyst, if you're a product manager, for example, or an engineer or an investor. And so in my own use case, it's mostly on investor or trader side. And that's also about, you know, 70% of our customers. So if you're, for example, going to acquire a token or like change your position in a token, either invest more into token or sell off some typically you might want to look at what are some entities that like have been accumulating this token lately, like the last 24 hours, the last seven days or 30 days, have they changed their positions? That's one thing that people very often look at.

Alex (00:18:15):
And so sometimes actually accumulation on its own will be a buy signal for some people depending on their strategy, because it indicates that they know something that the rest of the market does not know. So that's, that's kind of one, one use case. So the way I think about it, conceptually it's really, are you doing due diligence on a specific asset or are you doing discovery of new assets? And so in due diligence, you know, that would be an example of what I just said. You're looking at a specific asset, but in discovery you could look at, you know, what are some of the smartest wallets, you know? Cause you can actually think of them as entities in it. Even if you don't know who they are, what are some of the smartest wallets doing on chain? Are they yield farming from a specific project?

Alex (00:18:59):
Are they accumulating a token and so on? So that's, that's another case where with this yield farming craze, we've had many people have been looking at who are actually the yield farmers in specific contracts. And then they maybe, you know, maybe mistakenly, but sometimes they use that as a proxy for risk and they say, Oh, you know, this big fund is yield farming. This they're actually the biggest farmer of this contract. I will assume that they have done the risk due diligence and I will just ape in after them, which is, you know, sometimes it might work as a heuristic, but it's definitely not a hundred percent safe.

Anna (00:19:36):
And we heard about that in the SushiSwap saga episode that we did a few weeks ago where we we've heard about people kind of like, Tarun, I think you had an example of people pinging you... This is the same...

Alex (00:19:50):
We should find a code name.

Tarun (00:19:51):
Everything boils down to, to SBF lately. And so I think the analytics have been quite interesting in that once people started realizing, you know, I think in the normal markets, if like a whale influences the market and people like want to kind of like dox the whale because they like affected microstructure. They'll do things like call Bloomberg and tell Bloomberg like, "Hey, like there's this market manipulator" or like, you know, call the SEC or, you know, kind of make some noise that way. But the life cycle is so much faster encrypted, right? Like I feel like we went from like three weeks ago with sushiswap to like now SBF is a villain. He went from like the hero that saves SushiSwap to like he's the villain and everyone is looking at all of Alameda's on chain transactions.

Alex (00:20:41):
I mean, I think if it's a really interesting case as well, where you start thinking about how crypto is kind of a new paradigm, especially within regards to privacy, you know, reputation, identity and so on, because I think some entities will just choose to embrace the transparency of blockchain and you know, it makes a lot of sense that you'd want to increase privacy for individuals and even, you know, funds in many cases. But I think also obviously Alameda, I guess we can just say, you also said SBF before, so it's not, it's pretty clear.

Tarun (00:21:13):
Yeah, no, we've already, we've already, we, I was just saying.

Alex (00:21:16):
And so, and so it's pretty clear that like they have a much larger brand now, like partially because of all this stuff going on onchain. So is that a good thing or a bad thing? I mean, traditionally, you know, "all PR is good PR", so I think there's kind of a discussion to be had on that aspect as well, where you might actually use the blockchain as like an like almost like a social media. You know, it's kind of more like Twitter than, than a banking system, which is kind of strange, but in many ways, that's how it is.

Anna (00:21:47):
You just mentioned something, you just mentioned this concept of privacy on like in these analytics systems and obviously analytics is the observation of kind of data and the connecting of points, right? It's like linking things to one another. Privacy, this was a topic that we definitely want to cover on this podcast. How do you approach privacy? Because we talked about the previous analytics companies in their case, it was often about identifying users and reporting them potentially to an external party or even the example you just gave, like in DeFi, are you still trying to identify the users or not?

Alex (00:22:24):
It's a great question. So first of all, there are some regulations here, right? So even disregarding what we personally believe about something, there are regulations, they have to follow like GDPR, for example, in Europe. So we don't track data on individuals, for example. So we don't add labels for individuals with the exception of people who have kind of where they have placed information in a public domain already, for example, you know, my name is so and so, and I'm an, I'm a multisig signer of project ABC, that would be a scenario where we're comfortable placing the label. And if that entity comes to us and says, Hey, you should remove that. We will do it.

Anna (00:23:02):
Got it. Also, I guess if someone tweeted it, if they were like, this is my address, send me okay. Then you kind of make that association.

Alex (00:23:08):
Yeah. You know, again, assuming that, you know, that Twitter is still up, so it's in the public domain and so on. But really what we've found is like people aren't really interested in individuals as much. I mean, they're, they're interested in notable individuals of course, but then there's typically a corporate entity associated with it. And so we, when it comes to real world entities, we focus on, you know, projects, exchanges, funds, in some cases, mining pools, you know, these larger entities that are not tied to individuals, that's mostly what people care about. And so it's been not that difficult for us in that sense too, you know, find a limit between individuals and you know, all these different corporate entities.

Anna (00:23:51):
What you're saying here, are you saying that like on the individual level, you at least Nansen, doesn't have a focus on linking the individual account holder to an like a person, but you have identified, I guess, the corporate accounts.

Alex (00:24:06):
Yeah. And in many cases we have, I should also point out just for full transparency that when it comes to individuals, many people dox themselves onchain, like literally through DNS names. Right? Yeah. And you know, some people do that willingly and consciously other people probably aren't as aware of this. And I think sadly, a lot of people tend to think that if you can't see something on ether scan, then you can't see it on the blockchain, which is a really bad juristic because you obviously have the whole transactional history of, of the blockchain available for anyone. So if someone has the time, they can sit down and parse out all the onchain from the sections and identify exactly where all these ENS names are associated, like which wallets, which ones that own them in the past and so on. And so we do show ENS names in our platform.

Alex (00:24:56):
And sometimes that can be quite revealing. Of course, there's no guarantee that if someone buys, you know, AlexSvanevik.ETH that's actually me, but that's what it says on chain. So we don't censor ENS names as such, like we knew we needed to display them in the platform, but really the focus for us, it's not on individuals and we don't add any labels. I think it is also important to distinguish between the collection of data and the display of data. So in the case of ENS, that's not really collection of data. We're just displaying the data. That's already on chain. If we were collecting data on individuals like, Hey, you know, we found this information, someone told this person owns this wallet, you know, and if we put that in the platform, that's a very different use case from a privacy perspective and a data perspective.

Anna (00:25:45):
It's funny, like when you talk about kind of an analytics platform, I often, for some reason, for me to be honest, I haven't used, I haven't used one. I haven't used anything past a block explorer. And I think of a block explorer as being sort of like the first level of analytics. But I mean, all it is is really just like, here's some very basic data pulled from the blockchain displayed for you. This is what this, user's account number is called. Here's the amount in it. Here's the transactions. But actually what you just said there, you hinted at something like what isn't displayed in the block explorer.

Alex (00:26:20):
Oh yeah. I mean, there's lots of stuff that's not displayed. Right. So first of all, if you think about a product like Etherscan, which is great when I use it every day, I love it. I think it has different goals than most analytics products and in particular Nansen has very different goals. So first of all, Etherscan seems to me to try to be more of like a registry or a reference and it intends to be less opinionated. And that means it doesn't sort of favor the display of certain things over other things. And if you go to a specific address and you look just visually, you'll see all the transactions listed chronologically, right? There's no kind of visual emphasis on any particular transactions. And so Nansen is a bit different because we give ourselves kind of the Liberty to be opinionated about how we display data.

Alex (00:27:09):
And so we will typically make more emphasis on like large transfers, larger transactions because we imagine that investors and sometimes product teams are actually more interested in that. So that's one thing that is kind of, you know, there's a difference between the more kind of aggregated analytics and opinionated analytics. They can call it that versus the just blockchain explorers. And so we don't really have the goal of displaying everything that happens on chain in that sense with something like Etherscan has a lot of stuff you don't see. The ENS names, you have to kind of go to a separate section and look that up. You don't necessarily see past ENS names and I'm not sure why they kind of haven't focused that much on having ENS names in the main view. It could be because they are, I don't know thinking about the privacy concerns, but in my personal opinion, it's better to make people aware that these ENS names are actually there on the surface. Yeah. And so it's funny because I've seen after Nansen started getting popular. I think a lot of people have actually become more privacy aware, which is, which is a good thing in my opinion, because we are not out there to track down individuals who, you know, hold some certain amount of crypto in their wallet. That's not really the focus and people generally speaking don't really care about that.

Anna (00:28:32):
In your tool, do you also show all of the links, like every account and the links that that account would have to other accounts? Like if, you know, so you do you do that.

Alex (00:28:40):
So we do have, for example, a dashboard that's called the wallet profiler. And so you input an address and you see all the neighboring wallets that this one has transacted with in order to help you understand the context of what this wallet is. And it's, it can often be quite helpful to understand, you know, what are the, the different DeFi protocols, for example, that have interacted directly with a smart contract, what are the main entities in terms of exchanges or even funds or market makers like we talked about. And so, yeah, you do see that definitely that's, that's one of the more popular dashboards in the platform.

Anna (00:29:15):
I like that idea though, that like by looking at, and I actually, I would, I would say that in the web2 world with my old startup, when I learned about, you know, online analytics or Google analytics, like obviously like my realization of what was being tracked, my eyes were open to that. All of a sudden I was like, wow, I have to think about my own behavior. And you know, I, all of a sudden understood how my privacy would also be impacted by this. I haven't yet looked at anything on the crypto side. So I think it sounds like it's a good time to do it. And maybe this is a good recommendation to people because to better understand what it actually is. You'd have to actually see that displayed in a way, like what you've done, what your actual history looks like for real.

Alex (00:29:57):
Yeah. So one idea we've played around with, which I think we might do is to just give people a simple tool where they can actually just look up their own address and see like, Hey, here's what you can actually see about this address. And of course, you know, our product is a paid product, so we can't just let them do that for any address. But you could imagine that they just signed a message proving that they actually own that wallet. And because of that, they are allowed to see the transactions on their own wallet, which would be kind of a privacy tool that might help people get more aware of what's actually possible to see

Anna (00:30:30):
Totally. That would be actually amazing to have like a self reporting privacy tool for, I mean, DeFi general, blockchain. Like that would be so useful.

Tarun (00:30:41):
I, yeah, this is high, high up on our list.

Anna (00:30:43):
Is there anything like that?

Alex (00:30:45):
Not that I'm aware of. I mean, they, I don't know if Tarun has seen anything I'm not aware of anything, so.

Tarun (00:30:50):
Nope. Yeah. I haven't seen anything quite yet. Although I feel like that is a direction that you we're inevitably headed towards, especially I think as the projects themselves want better analytics on under users, I feel like that's going to be a market for us, like pushes this.

Alex (00:31:07):
Yep. So, you know, going a bit back to this discussion about privacy identity and so on, we do already have some examples of people that have like an onchain reputation. So there are ENS names that aren't, they aren't widely known who they are in the real world, the most famous one or the one that I follow the most is called newbie.ETH.

Alex (00:31:29):
And this entity did some incredible trades and some activity and like early yield farms. And I know for a fact that a lot of, a lot of people follow this account, but the funny thing is like, no.

Tarun (00:31:41):
I am guilty. Of following this person

Alex (00:31:44):
Yeah. And I'm sure some people know like, you know who this person is, but I don't know, but I still follow them. Right. Because, so, so in fact their address, is actually an asset in itself. You know, it's valuable if a token project has newbie.Eth on their top 20 list, as one of their holders, like I'm going to be looking at that project. It's just, that's just kind of how it is. So I think it's almost like a third face, you know, internet in the early days everyone was using like avatars, nicknames. We didn't really know who people were web 2.0, you had kind of the Facebook era where, you know, everyone had their full name on the internet and everyone knew who you were and you logged in with your Facebook and everything. Web3 might be kind of a return to the roots. And that sense where you can have a pseudonymous presence online and people don't know who you are, but you can still build up kind of like an asset on its own.

Anna (00:32:41):
For better or worse as we've seen recently. Exactly. Sometimes you can build a quick reputation in a month or two and then do a massive, a rug pull.

Tarun (00:32:53):
Yes. Right. But now everyone can watch you and we'll hold it against you in the future. I think there's actually this interesting thing of like an on chain analytics, forces, everything to be an iterated game. You can't just kind of like win a one shot game where you're like, Oh, I'm just gonna fuck over everyone. And like take all the rewards. Well, if, if there's some sense of on chain identity and you don't do a very good job of hiding it, which is expensive in some ways, right. You actually might not be able to win in whatever crypto game you're playing. You know, everyone knows and will start suddenly being able to turn against you. And so...

Anna (00:33:30):
That's interesting. I'm just thinking about that example of like, I don't know, like some cartoon character type avatar, Twitter personality, but like there are addresses probably linked to like anyone who's in crypto and on Twitter, they may have at some point like shared that link. And then even if they try to kind of move away from that avatar, they try to create a new identity. Let's say because all of it's sort of cartoony and fake anyway, it's not directly linked to them. They still have to use the same crypto accounts. They still have to use, like there could still be some sort of link between them there.

Alex (00:34:03):
Yeah. In, in you know, if you don't have good OPSEC, I guess. I mean, yeah. You could just create a new address and funded from a centralized exchange and then it's quite hard to track it, to be honest

Anna (00:34:13):
From the outside they exchange notes.

Alex (00:34:15):
Exactly. Exactly. So it's kind of hard to do it fully trustly. I mean, you, you do have like tornado cash and so on. And so some promising early projects in practice probably easiest now is just you know, fund an address from Coinbase or something like that. If you just want to be pragmatic. And that's what a lot of people do.

Tarun (00:34:36):
In a future where there are more private blockchains, how do you view the role of analytics and how do you view analytics evolving? If say like we have a world where privacy focused blockchains takeover, or like even like Layer2 are private. So you only get some, some data on the main chain. How do you view the evolution of privacy and analytics? Cause I feel like just like in the normal markets, there's this cat and mouse game of like everyone knew they're being watched.

Tarun (00:35:03):
They moved to something they thought was more private than they just found out that the new entity fucked them instead of the old entity. And, you know, but I feel like in crypto we actually have a chance for like separating church and state in terms of privacy and transactability. So, you know, we're not there yet, but you know, I'd love to hear what you think about kind of how it looks interacts with them.

Alex (00:35:24):
I guess the first thing I think about because, you know, we run a business around analytics. The first thing I think about is like, if you see either certain privacy chains or privacy technologies or Layer2, like how will this impact our ability to do analytics. And just from the perspective of our team, three different scenarios, either it's impossible. Like either there's some technology that just doesn't make it possible to do certain types of analytics, which seems kind of, I don't exactly see how that would play out in practice because you know, there are going to be some kind of traces of things happening, but it might not be as kind of parsable as it is today. So that's one outcome. Like it's just impossible and then it's gonna be impossible for everybody and not just for us. Right. And then at the other extreme, it's going to be relatively easy to still do analytics, which, you know, let's say you have some new chain that gets very popular and it's similar to Ethereum 1.0 today. So it's still going to be quite easy for us to do it. And in that case, you know, we're going to be able to do it since we've done it with Ethereum 1.0, and then there's kind of this step in the middle where it's going to be just a lot harder.

Alex (00:36:33):
And in that scenario, I think analytics companies are going to become even more important because it's harder to do it on your own. And so you're going to have to rely on analytics companies. So that's almost like the best case scenario in a way, if it gets, if it's not impossible, it's just really hard. I think that's a scenario where analytics companies will actually thrive ironically, perhaps because a lot of people are kind of, you know, they ask us like, what are you going to do in that scenario? But that's actually the best scenario for it. I think strategically for analytics companies. With L2, so taking a step back from like our business, which is obviously top of mind for me, I think with L2 you might see more fragmentation. Right? So that means you might have more niche players focusing on specific L2 technologies or basically, you know, second layers.

Alex (00:37:21):
And so you might see that there's a standard that evolves for like gaming, for example, and then you'll have gaming analytics companies or blockchain gaming analytics companies pop up. That could be one scenario and with other chains, I think that's, you know, that's a tricky one. If you just have like other chains than Ethereum get to the same level of ecosystem activity, there's not much evidence that we have any, I think so far, but know if that happens, we'll probably see either, you know, some of these companies move over and actually support other chains in our case, our technology, it doesn't really rely on Ethereum specifically. Like we have an obstruction layer above that, that we can use for other chains as well in most cases. But you might see that a new wave of, you know, if it's near protocol analytics or, you know, if it's Polkadot analytics or whatever it is.

Alex (00:38:14):
And yeah. And in the case of privacy chains personally, I don't, I, you know, we haven't seen that much evidence that people are using privacy features, even the ones that exist today, you know, so far. Yeah. So far. So I guess the verdict is still out on that if it's gonna really take off and also, you know, the regulatory aspects might become a bit trickier. Yeah. Yeah. So there's different things there. Just one last thing I would add on it is that I think it goes back to this transparency versus privacy dichotomy. And some entities will obviously want to prefer transparency. Like if, if you're a CeFi protocol, like a lending protocol, I think it's actually competitive advantage. If you can show a transparency, like when you have certain funds, like I not sure I would deposit funds into say like Nexo or Celsius or BlockFi if like their whole stack was kind of privacy first oriented where you don't really know what's happening with the funds. I think I'd be more comfortable depositing funds into a CFI protocol, if there is a degree of transparency. That's just me personally. And I think some people might feel the same way, but you know, there, there is that kind of dilemma between transparency and privacy,

Anna (00:39:32):
Actually a solution to what you just described. I mean, there are zero knowledge proof systems and concepts already in place, which suggests this idea of selective disclosure, where you would be able to prove, you know, for sure that the funds are there without revealing how much is there exactly. Or without any sort of information leakage. That's where like, I mean, I love this idea of it doesn't have to be privacy or transparency, but you could actually somehow find a way to have both.

Alex (00:39:58):
Yeah, that's really cool. But, but like in practice, just kind of riffing off of that in practice, like how would a platform like Nexo do that without having a third party that almost access, like the trusted party that I can actually in an easy way show to other people that they actually hold the funds...

Anna (00:40:14):
Where like, if it's, especially if it's numbers you could actually use as ZKP to prove that it reaches a certain threshold and all you would get is the answer would be, I mean, I'm simplifying like crazy. But the output would be something like, does this CeFi entity actually have it? The output would be true, right? That's like a very simple way that I've understood this idea of selection of disclosure using ZKPS. And it works best when there's a value, like a numerical value or something very mathematical it's less easy when it's something like you would need some third party checking your ID, like say it was like your age. It's like you still, at this stage in time would need some entity to say, yes, that is their age like the I don't know the DMV or something. And then you could put that on chain and then you could prove it using ZKPS, but it's, you're still, you have this sort of arbiter the same real-world through third party entity step that you described. But if it's numbers, you don't need that so much, but I don't know exactly how they would implement this, but I do hope to actually see that combination someday.

Alex (00:41:21):
Yeah. A hundred percent. I mean, that sounds really, really interesting.

Tarun (00:41:25):
They still had not been good designs for financial exchange mechanisms that user knowledge proofs, for instance, like something Uniswap, like is fraught with a bunch of difficulties because you have to have some things that are global public state that everyone agrees on. And some things that are private state, like, Hey, I made a trade, but somehow not leak information in terms of like the timing of when you send orders and things like that. So I think there's still a huge design space that people who work on ZK piece, I think is basically more focused on, you know, simple circuits and things like that. But I think for a lot of these DeFi applications where we don't have any of the economic mechanisms built that are compatible with privacy, to some extent. So...

Anna (00:42:10):
Yeah, I think this, I mean, this is a topic I know Tarun you and I have talked about this, but this is a topic we want to cover eventually, which is that of privacy and DeFi. Like whether they work together, I think when it comes to the question of analytics, cause we'll bring that, we'll bring it back then to analytics. Like it's almost like, I think you're kind of correct in saying that like right now they're the privacy chains, maybe aren't useful enough or these even privacy Layer2, they're not seeing the volume that you would necessarily like need dedicated analytics to deal with yet, but they could. And yeah, I kind of want to, I want to hear like, even a little bit more about what you were thinking with the, like, how would an L2 analytics, like, how would it, how would you even track that right now? This is actually kind of taking it outside of privacy, but going back to the sharding or heterogeneous shards or these other kind of, you know, interconnected interoperable systems or Layer 2s, how would you even approach something like that?

Alex (00:43:10):
I think at least one way you might do it is to simply rely on a certain project own APIs. Let's say, you're you got a gaming oriented L2, they might actually be somehow tracking transactions that they place, on a Layer 2 cause this doesn't necessarily have to be privacy oriented as such. And so we can still parse out, you know, the data from either their API so that they provide to us. Or in some cases you could read it directly off a chain of if you're a validator or something like that on, on that Layer 2 network. In the case of Zero Knowledge and so on, I'm definitely not an expert, so we should get Evgeny or something back on this podcast and dive into that. I think, you know, one year from now we might have some more real world experience with some of these L2 solutions and so on.

Anna (00:43:56):
That's true. I mean, it is. So I know I'm sort of throwing this question to you. Like you're supposed to have an answer right now, but like seriously, like I used an L2, I think for the first time, a few weeks ago with zkSync. I don't know if I'd used it before. So with, with the Gitcoin kind of grants and stuff like that, where I actually like looked at a block Explorer on an L2, and like proactively interacted with it. So it's pretty fresh, you know, maybe I'm late to the party, but like still, I think I'm more probably like a, I always call myself a "late early adopter", so if I, if I'm there, we're about to not be early adoption time, but I just tap into the end.

Tarun (00:44:36):
I don't think you're, you're quite late. I don't even think the UX on Layer 2 is the solidified. Yeah, I agree. It seems, it seems like we have a new, most of them are still on testnet right. Hasn't really been one that's been like, Hey, we're in production and the interfaces have changed a bunch and the validators sets have changed. So I think we're, this is like 2012 in Bitcoin for, for L2s, maybe even 2011 now.

Alex (00:45:03):
Yeah. I mean, it's pretty awkward to interact with these right. Like you have to, you know, put in like custom RPC or like all these different things and Metamask or whatever you're using. It's, it's definitely not a delightful user experience.

Anna (00:45:15):
It sounds like for both the topic of privacy and the topic for sort of L2, this is like a moving space. And so in my opinion, it's something to like keep an eye on and it's definitely opened up some new questions for me and something to explore.

Tarun (00:45:29):
Yeah. So, I mean, I think one thing that's interesting is, you know, we've talked a lot about trading and keeping track of users who are in trading or like monitoring different users and how there's actually in this crypto world, unlike the normal world, there's this interaction where the analytics influences the market and the market influences the analytics and you get this feedback loop. On the other hand, you have positive feedback loops too, right? So there's, governance where, you know, in order to make governance decisions, people need to have really good data to justify numbers that they're going to pick or to justify why certain types of entities should be excluded from rewards. I'd say the Uniswap's Dharma Airdrop, has been very controversial because of that or potential airdrop. But, you know, I think there's also these kinds of positive sum games, not just zero sum games like trading where there's a huge benefit from studying these things.

Tarun (00:46:26):
And so I'd be curious to learn what are some of your most surprising learnings from kind of some of your research, especially when a lot of it might not be the trading use case like there, because, you know, I think there's starting to be a lot of other use cases and it'd be great to hear like what some surprising things you found there.

Alex (00:46:44):
So governance is definitely an interesting area, but we did some work with Aragon where we looked at participation rates, for example, in AGP votes. And, you know, that was an interesting process because we could actually see, for example, you know, what you might describe as insider voting,we're basically you have wallets that are a small degree away from certain specific, like a team multisig or, you know, a token sale contract or things like that. And that was actually something that Aragon like asked us to find for them, right. Just because it's something that the community said, no, if you have examples of that, and it's not necessarily, it's not like, you know, insider voting sounds a bit harsh, but it's really just people who are, you know, central in the project in some cases. And they were, of course, totally allowed to participate in government as well. And so we did an interesting project where we kind of reviewed every single AGP vote that's ever taken place on Aragon. And it was kind of interesting to see, for example, what are the different votes that drive more engagement? So like voter engagement,uin itself is an interesting area. I wouldn't like single out, you know, a specific vote or anything like that in the, in the case of Aragon. But it was some, you saw some clear trends that certain votes that related to like other projects, for example, like, should you fund another project? Those votes had high engagement and then other votes that were more kind of admin oriented, you know, they were basically like no engagement at all. And it was just like a couple of, call it central key players in the ecosystem that voted only, more out of chore if anything.

Alex (00:48:29):
So I think that whole area is really, really interesting. And as you say, I think it's positive some to understand governance processes better. And I think most people would agree that the kind of governance experience today is not that great. Like there's definitely room for improvement. On that there's a lot of voter apathy, even just voting in itself, like the mechanics of it is costly with gas fees and so on. It's been a bit better now with snapshot.page and so on, but like definitely the whole governance area I think is really, really interesting. And I personally think that you kind of need to have the kind of entity contextual layer on governance. Maybe not necessarily knowing who the real world entity is, but having like a persistent view on like who this user or who this like yeah, exactly.

Anna (00:49:20):
Even if it's not like a person and that reputation has to hold value, maybe even outside of just this governance thing where it's like, if you, cause what you'd, wouldn't want someone to build up reputation in this tiny little kind of similar to the example I gave before, like where it's still only lives in this little bubble. So you have a good reputation there, but there's no stake. Really. There's nothing to be lost if on one decision, they forfeit that reputation, but they, you know, make off with a ton of money then maybe for them on that one in that case, it doesn't matter. So like some link between that inner reputation in the external world, you still need, even if it's sort of pseudonymous or not connected to a person.

Alex (00:49:59):
Exactly. And again, you know, transparency can be quite useful here as well. Like if, you know, the certain positions of some entity with regards to which other assets they hold that can obviously influence how they vote. And so just being aware of that can be useful. Of course, you know, you can hide away assets and other wallets as well, but that's yeah. I mean, I totally totally agree with that assessment. So I think governance is a really interesting and like comp it's a complex where, you know, things like Gauntlet, for example, are going to be extremely important in making governance better. And you had a good example of this recently with unit right? Tarun.

Tarun (00:50:39):
Yeah. I mean, I think one of the things, especially as I've been doing more research and trying to make this more, understand how people who are interacting with the system should really think about parameters and how people should write proposals is that in crypto, the wealth of data is actually a lot better than what you have in the normal world, in terms of governance, where, you know, you're trusting some type of lobbyists data collection and analysis for, you know, for better or worse. Like if you've ever interacted with government agencies, they're always trusting some random consultant who went and scraped data for them and gave them a .csv. And then unfortunately, Excel overflowed because they didn't have enough cells kind of, I don't know if you guys saw that the Excel bug with Britain, but basically Britain missed 1500 COVID cases because they had too many rows in an Excel sheet and it overflowed Excel, like Excel's limit.

Alex (00:51:41):
That's insane but also very unsurprising if you've worked with data in the real world, like this stuff happens. Right. Which is crazy.

Tarun (00:51:48):
Well, it's just funny that it's like sitting in like the entire countries COVID stuff is, their Database is an Excel sheet, right? So like, this is the type of stuff where like crypto and governance is it's way more transparent. Like why are there three people who are looking at this Excel sheet that contains the health of the country? But what I mean is like, when you make governance decisions in crypto and DeFi in particular, you just have so much more data to use, which is great, unlike the normal world, but you also have to figure out how to distill it to something that is a clean, easy to understand piece. And I think, yeah, that's where I see a huge amount of value in the analytics and for transparency kind of thing, because I think, you know, we've seen version 1.0 crypto where people kind of vote randomly, or they make random proposals, or it's like extremely interested party furthering a certain agenda, but there is, you know, there's going to be the more people that use these systems.

Tarun (00:52:48):
The more there'll be things filled out in the middle. And I think personally I've found using Nansen has been really, really useful for, because it looks at off chain entities and on chain entities and how they interact. Yeah. And I think that's the thing that, you know, I think the old school analytics companies like chain analysis, they only focus on that from the perspective of, Hey, is this person a terrorist or, you know, money-laundering or whatever. But I actually think when it comes to behavior, it's more complex and you really do need to understand smart contract state, to be able to say something useful about why someone should vote on your proposal.

Alex (00:53:26):
Yeah. Just to, just to kind of add to what you said. I mean, this was kind of the main reason why I started working full time and in crypto that there is this abundance of data, but it's still quite immature in the sense of, you know, how you use it for analytics. And so, you know, if there's anyone listening is like a data scientist or data analyst or data engineer, I mean, I really strongly believe that this is the most exciting industry to work in if you have a skill set, because there is so much data at your fingertips and there's still tons of work to do on the analytic side. So I can't really imagine any better industry to work in. If you, if you are interested in data and analytics, it's just even blockchains themselves are actually really neatly structured. So you have really high data quality. You don't have these kinds of Excel issues that Tarun just mentioned. It's just a really great industry to work in. If you like analytics.

Tarun (00:54:20):
The funny thing about that is that like, you know, most people I know who are in kind of data science, statistics and machine learning, like they, they still have this aversion to crypto like, Oh, it's just like scammers. And it's like, everyone just Ponzi, scheming each other. Why would I ever want to work in that? Like I remember when I left, I got a lot of ostracism from like my friends who were in sort of more like statistics and data science and stuff. They're like, Oh, like why, why would you like leave the clean stuff for this kind of scammy area? And I think what we've seen is that, well, first of all, COVID has shown that everything is a scam by looking at the stock market. But secondly, you know, I think data being used for bad has finally come to light to a larger swath of the population that like, it's, there's no free lunch to this kind of like platform based data science and you kind of towing the line between transparency and privacy is probably the future, but yeah, crypto still has a very bad reputation amongst people.

Alex (00:55:26):
It does. I mean, so just to draw an analogy, it's, it's maybe not a great analogy, but when I started studying machine learning in like 2007, 2008, I remember talking to people who were in kind of just computer science in general, about how, how exciting like AI and machine learning is, you know, back then, it wasn't really clear that machine learning was going to have that big of an impact as it has had in the last decade. And so a lot of people just dismissed it and were like, yeah, machine learning. That sounds fancy, but it's really, really boring to work with in practice. It's not interesting at all. A lot of people actually told me that and you know, you might see the same thing now. Like people tend to have a certain skepticism towards new technologies. They don't believe in it before they actually see what's, what's possible to do with a new technology.

Alex (00:56:15):
And crypto has like twice that challenge because you also have like the, the scams and all that stuff in. But you know, I think we already have some amazing examples of what you can do with DeFi and like the one, some of the first projects, I show it to people are compound and doing a swap, which actually solve real world problems. And they're easy to use. They're fully decentralized with regards to governance and so on. So I just hope, you know, I think we're going to get more great examples like that. That's overtime. We're going to draw people in, even if they're a bit skeptical right now.

Anna (00:56:52):
Cool. I wanted to ask you, so you like on the, on the website, there is this A.I. Focus. How do you actually use A.I.? Because we haven't spoken about that at all. So far. Like where is the A.I. Component in all of this?

Alex (00:57:05):
Yeah, so I think, you know, A.I. At this point is kind of a buzz word, but the machine learning component is really, you know, that's how I would describe it. So when it comes to, for example, scoring the probability of a wallet, having a certain label, that's one example of how you'd use machine learning. So for example, what's the probability that any random wallet belongs to Coinbase. So it's controlled by Coinbase. And then you can think of this in like basic terms or probabilistic terms. And you can add evidence that you have from the onchain data. So for example, does this wallet have different tokens in the same wallet or does it just have one type and does it have, you know, a fixed size in batches of transactions coming into it? What's the gas price that is used on the transactions? Does it have this certain kind of graph features in the transactional network?

Alex (00:58:04):
And all of these become features that you can use in order to make a probabilistic assessment of whether or not it belongs to Coinbase? So that's how we use it in practice. It's not really A.I. In the sense that we're kind of just, you know, speaking to like a robot or something that then we're asking them about addresses, but it is the machine learning component and machine learning is basically what powers all of A.I. Today. Cool.

Tarun (00:58:28):
Uh what would it take to bring people to the space kind of in the same way you just talked about the transition from A.I. Kind of being like pie in the sky to like every under CS undergrad is trying to do it, what type of things would be needed for that transition? And, you know, personally, I think having really good analytics tools like someone has to make PyTorch or TensorFlow for this stuff.

Alex (00:58:53):
Right? Yep.

Tarun (00:58:53):
And like, that's kind of what we're all doing on, like, in some sense, but like what are kind of the tools that you think that would kind of move us in that direction too? Like every, you know, smart undergrad is like: Hey, I want to work on this.

Alex (00:59:09):
Yeah. I'd say maybe two things: The first one is I do think that we need a bit of a culture change in crypto in the sense that we tend to be very inward facing and actually not that outward facing basically it's a full time job in itself to keep track of what happens in DeFi. And that's not, that's a good thing, but it also makes it not that easy for newcomers to have to, you know, understand these six layers before you get to the seventh layer of the product that you're using. So I think just generally a mindset shift where you try to actually make it a bit more welcoming and just easier for new people to participate in the space.

Alex (00:59:48):
I think that that's a good thing. There's a lot of people doing great work on that front, but I think we could be even better. And then the second is, you know, since you mentioned AI and machine learning, as the analogy, there are some things that actually popular popularized machine learning heavily. So Kaggle is probably a good example where you had machine learning competitions where a lot of people actually started doing machine learning on Kaggle because you had this incredible forum of people sharing how they built certain, certain machine learning models and how they were able to win competitions. And so having kind of either competitions or more like, you know, hackathons and open forums for people to participate, I think that's a, that's another area where you can learn from machine learning and AI. And of course the fact that Kaggle also had like a financial incentive, you could win. Lots of money was, was really good. They also did great work on the PR side, right. They were able to get some incredible stories posted about some person from some field that was able to make a massive breakthrough in another field just because they know how to do machine learning. Right. That's like incredible marketing for machine learning. So I don't have the exact solution, but something along the lines of Kaggle for crypto, I think would be really cool,

Tarun (01:01:07):
Dark forest. I think the games are kind of the, those are actually really good. That's a really good point. Like being able to use X for solving problem and Y is usually ignored in crypto has, has spent most of its time being like, Hey, how do we take X from somewhere else and apply it to crypto. But yeah, I guess that hasn't really made any other field kind of improve.

Alex (01:01:32):
Yes. I think you're right. Gaming is a really exciting area for this. So another example that I love is Axie Infinity, which is this NFT based game. And they had these incredible stories of how like people in the Philippines and a couple of other countries were playing these games and they're actually making money off of them because there's like, you get dire rewards from playing the game and you can actually sell certain, like potions that you earn from playing the game and you can sell them on uniswap, you know, to the open market, which is really cool. And then in addition now with the uniswap airdrop that happened, like there were people in the Philippines and Vietnam, like all over the world who just overnight received like $2,000 in Uni tokens, which was this like incredible thing on its own. That's not really, you know, addressing necessarily like that technical crowd, but it is definitely a way to make crypto more mainstream and to get those incredible stories out there on how people can use it and actually benefit from it directly. Cool.

Anna (01:02:33):
So now, if anyone wants to learn more about Nansen specifically, what should they do?

Alex (01:02:37):
Go to nansen.ai, or you can also follow me on Twitter, @ASvanevik, it is my handle.

Tarun (01:02:46):
And a one thing is anytime if you're on Twitter, if you see "Medium DEX Trader" as a meme, that is, that is a, one of their labels that became like kind of a huge meme over the last few weeks. So if you see that in the screenshot, you'll know that's cool.

Alex (01:03:01):
We might, we might have some Medium DEX Trader mugs coming up at some point. Yeah.

Anna (01:03:05):
Well, thanks so much for coming on the show and talking to us about analytics, crypto analytics, DeFi analytics, and, and all of the other topics we covered. I feel like from what we talked about, I've already had in my mind, like the idea for three new episodes to kind of go a little deeper into some of those directions. So it's been cool. Thanks a lot.

Alex (01:03:24):
Thanks for having me really fun being here.

Tarun (01:03:26):
Thanks for coming on Alex.

Anna (01:03:27):
Yeah. And Tarun, thank you for co-hosting again.

Tarun (01:03:29):
Hey. Yeah. Thanks for having me.

Anna (01:03:31):
And to our listeners. Thanks for listening.