Future of Social Cryptocurrency in Web 3 with Pi Network's Co-Founders, Dr. Nicolas Kokkalis and Dr. Chengdiao Fan

Welcome to the Index Podcast hosted by Alex Kahaya. Plug in as we explore new frontiers with Web 3 and the decentralized future.

Hey everyone and welcome to the Index brought to you by the Graph, where we talk with the entrepreneurs building the next wave of the internet. I'm your host, alex Kahaya. In today's episode of the Index, I'm thrilled to welcome Nicholas Cacolus and Cheng Zhao, the co-founders of Pi Network, the first digital currency that can be mined on your phone. We will learn all about how Pi differs from other crypto projects, their approach and the utility that they're building. Welcome to the show. Thanks for being here.

Thank you for having us, alex. Thanks Alex.

So I would just first love to learn more about you guys. How did you meet? How did you get the idea for Pi? What is the genesis story in your backgrounds?

Pi Network is a combination of our academic and also personal interests. Nicholas and I got to know each other a long time before we built Pi, so he is always a hardcore computer scientist, from hardware to theory or distributed system, which is a precursor to blockchain and later human computer interaction. I am a social scientist. During my second half of my PhD at Stanford, I also got interested into human computer interaction, so that's where Nicholas and I collaborated many years on multiple projects in social computing, crowdsourcing, building large scale systems, design systems for scaling, human communication and productivity. We wrote a few papers together and then we both are interested in entrepreneurship. So we combine all this interest together. Pi combines all this interest I mentioned above and to solve human interaction, facilitate large scale participation and collaboration to create something larger than just the people combined collectively. And then we think blockchain is a very powerful technology to facilitate that and we look into the solution in that space. And actually Nicholas was even teaching a class on decentralized applications, on blockchain, and through that experience we discover certain issues of the industry at the time.

Especially through that experience, we were very frustrated with the current limitations and implementations of the crypto industry at the moment. We initially tried fixing the user experience problems that some other platforms had. Soon, we realized that without actually building our own platform, we would not be able to create a truly decentralized and truly accessible cryptocurrency for people to use. So that's because, fundamentally, people would not have been able to mine it themselves without upfront financial costs. We ended up designing Pi to address all these issues and we aimed to have it have accessibility and wide distribution on a large social network that recognizes Pi and, this way, allows true utility to emerge.

Really interesting and it's all done on the phone. One of the things you guys are trying to accomplish is just lowering the barrier and to make it more accessible, and I think that's been the holy grail of crypto for a while. I remember I got back in the space in 2016, and that was the narrative then and it's still the narrative now. It's still a work in progress. There's a lot to be done, but I'm just genuinely curious to learn more about your approach to solving that problem. Let's go a layer deeper and talk through what it actually looks like, what Pi does and how it brings that user experience problem and accessibility problem fixes that.

Yeah, accessibility has multiple aspects. There is the financial barrier, there is knowledge barrier, there is also the easy to use or the technical, logistic barriers. So in order to achieve true accessibility, you need to solve all these barriers in order to break down all the barriers and have a free flow of user experience, and that's what Pi is trying to address those issues so, specifically, have a phone that is accessible. Almost these days, everyone can have a phone. And also, accessibility is about equal opportunity. So it's free, everyone can just go to the app store and download and then easy to understand, meaning that people don't have to have a CS degree in order to know how to run a node to mine, but they can use a phone and an app and building the user experience easy enough for them to follow the education of the whole process of what is mining, how you can mine. This is also important.

I would love to dig deeper into the consensus or just how that works From a mining perspective. That is definitely one of the biggest things, especially for early adopters, that you miss out on If you're not highly skilled DevOps engineer. In most cases you cannot participate at that level in the earliest days. A lot of that has to do with the software, the validator software, and how complex the algorithms are or whatever to validate and to set up that validator. How does it work in your system? It sounds like you get set up really easily, but under the hood. I'm curious to learn more about that.

The essence is a different type of consensus algorithm In terms of how it works. If you compare it to Bitcoin, everyone knows Bitcoin uses proof of work. Let's compare it to Bitcoin. Like Bitcoin, Pi is also running on its own blockchain, so it's a later one solution. We are not running on other people's blockchain. Unlike Bitcoin, we use different types of consensus algorithms. Bitcoin uses proof of work. The requirement for that consensus to run is contributors or miners contributing energy and electricity. In return, they receive Bitcoin rewards. In Pi's case, we use federated bisanding agreements. In Pi's specific blockchain, it actually relies on a trust graph. A trust graph is an essential component feed into the blockchain for it to reach consensus. The mobile miners are actually contributing to this trust graph that the blockchain will use in order to secure the ledger. Each miner, each mobile miner, actually are contributing by their trust circles we call it security circles, which is they appointing three to five people they trust. In aggregate, millions of people's trust circles forms this trust graph that will be used by the blockchain and to scale the mining in analogy is Pi's consensus algorithm requires the trust graph and contributors or miners contributing to the trust graph in return receive Pi rewards. This is the fundamental reason why Pi mining doesn't consume the level of energy required by traditional mining of cryptocurrency. At the same time, it's also allowing mobile miners to contribute to the blockchain without getting with obstruction also from the complications of the technical aspect of blockchain.

So I have five friends and I onboard them to the Pi network as part of my trust graph, to like, send them a link to the app. They download the app and they hit the button to run to contribute resources. How is the trust verified? I guess another way asks is like how would you game the system and how do you prevent that? So what is trust actually mean and how do you, at a scale? I kind of see it already If everyone's trust circles 30 people, that's still not big enough to be like the billions of people they can't possibly collude to corrupt the chain. I'm trying to wrap my head around, like what are the risks and tradeoffs? And like how does it actually work? How does the trust actually work?

Yeah, I can try to explain intuitively how this works. First of all, I want to clarify that the blockchain consensus algorithm is running on computer nodes and we'll have tens of thousands of these computer nodes already, even just on our testnet. So comparing the testnet of Pi with other mainets is a few or maybe orders of magnitude even bigger. Now the mobile miners. What they are contributing is those connections that you mentioned, the trust. So intuitively this helps isolate any potential. It's called Sibyl attackers. Sibyl attackers when someone is creating fake accounts to pretend he's more than one people on the network. So think of it intuitively if me, you and Cendiao are the only people on Pi network and no one else yet and we point at each other because we pulled out on the phone each other and selected them and we say that these are the people we trust. Now if a fourth person comes along and creates, even if they create a thousand fake accounts, the best they can do is they can have those accounts pointing at each other. But in order for us, for the three of us right now, to trust that, those fake accounts, we would need to be convinced to change the trust pointers that we have from each other to those fake accounts, and why would we do that? So now imagine take this idea to millions of people. You will understand how it works. In its essence, it's similar to the original Google page rank algorithm, where you have web pages on the internet and the true web page of MIT is pointing to the true web page of Stanford and it doesn't really matter in the end of the day, it's the authority that's built from backlinks is kind of the idea.

Yes, exactly.

Exactly, exactly. So you could have fake blogs pretending to be Stanford, but no one cares.

So let's use the web page example. Let's say I'm like a web page that's rank one on like a high traffic search. I have a lot of weight because I have a lot of trust. Like I have a thousand people pointing to me and you know those people are not fake accounts. Does that mean I'm mind more of the currency because I have more? Is it like proof of stake, kind of like that, or how does that work?

No, we didn't want to make it on the plutocratic side, so the mining rate is standard across everyone. There are certain bonuses depending on different contributions.

Basically the transparv give rise to the better formation of the corons, of the nodes, and the corons are basically a subset of the whole node network within which the nodes are messaging each other in terms of voting whether a transaction is true or not, and each time the corons are dynamically formed, so it's not just always that set of nodes. That's why it's to some extent more decentralized in a way, because it's a dynamic, corons is formed independently and within the forum, nodes message each other and reach consensus and the algorithm basically can ensure a finality to this transaction and get recorded on the ledger.

And so you mentioned that you have 10,000 of these nodes. Is that 10,000 people with their phones? That was my understanding, but is it different than that? Are they actual computers? They're phones in the sense that you're a computer.

Yeah, if you're referring to the blockchain running, to the devices that are running the consensus algorithm to produce blocks, these devices are only computers at the moment, not phones. We don't think that it would make sense for phones to be doing this kind of job and in fact, I mentioned that there are tens of thousands. I need to look a little more into giving you a more authoritative answer of how many of those we have, but some unofficial survey we did recently gave us more than 90,000 nodes on the network. These are the computer nodes.

Yeah, and that's just like laptops, like anybody's laptop could.

Yes, yes, so that essentially, if you look at this large number, even on just on the test net, it makes pi network one of the largest distributed systems on the planet. Not just a blockchain distributed system, any system that doesn't have a essentially centralized server. There's a lot of different kinds like file sharing, including distributed system.

And so you have a test net. You haven't launched mainnet yet. What's sort of the status of where you are in building, building all this out?

Our mainnet is live. We launched the mainnet. This goes to also your just now, your question about what we are different from other blockchains. We are taking very non consensus type of approach. Our mainnet is live and we actually created a special period is called enclose the network period. This is a special phase inside our mainnet where the mainnet is live but it artificially put a firewall against any external connectivity. This is to ensure that we can focus on building utilities within our ecosystem and we believe their utilities can be built in the enclosed environment. Of course, recognizing there are certain utility requires external connectivity, but we want to focus on building utilities within this enclosed phase in order to get the utility based ecosystem more mature before we open up to a open network where the firewall will be removed. In addition, during this phase, as you know, our network is not just a few thousand people or participants. We have millions of pioneers all over the world. So in order to onboard them, these millions, all into the network, it makes sense and more fair for majority of the network migrate to the mainnet in the enclosed environment before opening up. This is also another goal of the enclosed network, in addition to creating utility within the enclosed system, and this is also, you know, the answer to your question how is it so different from other crypto projects? I think this is an innovative phase we created.

Yeah, it's just kind of gated. I mean you let a bunch of people sign up and then you sort of gate the mainnet access. It's permissioned right now, but you will be an open, permissionless chain once you feel like you've kind of made it out of this period and you have enough utility.

Yeah, during this period, as we mentioned earlier, there is more than 45 million engaged members on Pi network. Right now. We call them miners. We are in the phase of completing, accomplishing KYC for them and transitioning them to the mainnet. The mainnet you see on the Block Explorer is the real mainnet. That will will continue when it's open. Right now, with KYC over 5 million people, and that's one of the goals to to make the migration.

Yeah, this choice of this strategy is also based on price philosophy of inclusivity, right. Our requirement of KYC is also making sure that all corners of the world regions are able to KYC and get on to our mainnet. This period supports this process.

I have so many questions. Just to take a step back, the utility you're building. Can you give me examples of what things you're building on top of Pi network and the examples plus the impact? You've obviously prioritized some things that are gating the launch of the permissionless network and so I'm kind of curious like what are things and what is the impact? You know, once you have these things, why do you need them to wait to be permissionless?

Yeah, so utility is decentralized efforts from the whole community, not just the core team. Of course, we play a big role inside that by providing the infrastructure that facilitates the decentralized efforts, such as providing the platform, the SDK and developer tools. However, the actual building and the holding on the real, true utility for human to solve human needs is a completely a collective effort from the community builders and also community users of those apps. We have ourselves built some apps, for example, just now I already mentioned the KYC app. The KYC app because, also following accessibility, we did an outsource to a third party, which usually charged at least $1 for doing a check right, so that actually we developed a native solution that not only solved the scalability and also wide distribution of the geographic locations, but also it doesn't require people to pay out of pocket cost. Actually, people, pioneers, actually pay high cryptocurrency for the labor or the work down by the crowdsourced human validators within the community to perfecting the validation process where the AI or machine automation lacks. Ai advance a lot, of course, but for identity verification there are cases that still needs human touch and that human touch was provided by the community itself. So actually, pioneers are working as validators, actually making pie from the pioneers who want to verify their identity. So this is a true utility using pie meaning that the pie is used to facilitate transaction between the labor of the validator, work of the validator and the result in actual substance, which is whether someone is verified or not.

So you guys built a custom KYC solution.

Yes, yes, we did. We have to, because even the market providers don't necessarily provide KYC in, for example, regions that are really small or have a smaller market. Because most of the KYC providers are profit driven, building the infrastructure for a specific small region really cost a lot for those businesses. But for pie, in order to really achieve inclusivity, we have to provide KYC solution to all regions where our pioneers exist. So we created a design innovatively, this design that's scalable to cover all geolocations while at the same time, not costing out of pocket.

So this is super interesting to me because and just for people listening KYC stands for Know your Customer and it's a really common thing you have to do in this space like anytime, just for like a lot of regulatory reasons. It's the best practice to KYC folks for things like grants, programs and kind of anytime you're dealing with tokens you're probably KYC'ing somebody, especially if it's a permission chain, like this is right now. So I see why you're having to do it, but building a totally custom one. That's a huge effort I've actually built, like integrated with providers multiple times to automate KYC for things like grant programs. They're checking OFAC databases, they're checking a whole bunch of different things to make sure that. I mean, it depends on the level of KYC you're trying to do, but making sure you're not on a list of countries where you're just not allowed to do business, like North Korea and stuff. What is the process under the hood, again, and what is running there in the app and how do people verify identity, especially this manual process? It sounds like with pioneers who are doing the work. How does it work? What are they looking into? What's automated, what's not? I'm just very curious to hear more about that.

Well, it starts as a typical KYC solution, what you would expect. So you pull your phone, you have to show liveness, turn your head, respond to certain prompts, you have to show your ID, and the ID is captured and then AI is reading it. You have to fill out certain information, which is pretty standard stuff that all the KYC solutions do. We have a back end layer which is processing all those things automatically, as much as it can and as expected. There are some cases where they in certain questions that the AI has or is uncertain about, and those questions need to be answered. Every KYC solution has to answer them with humans. So what we're doing at Pi is that we reward existing, selected, trustworthy KYC already KYC members of the Pi network to answer those questions. We also had to add an additional layer of security on this so that what they see is not the raw data of what someone updated. So, for example, if someone were to see, to want to verify the face of someone on an ID versus whatever we got from a selfie picture that they get, then our software is cutting out only the picture from the ID, is cutting out the picture from the selfie, putting the two together and asking a validator to say are these two pictures the same person? Another use case, and there's multiple of them, but it will be. The last example I give is that if you need to ask certain question about an ID to a human, we are. The software is covering with black reduction boxes all the information on the ID, that all the text that is not supposed to be seen, and it's also cutting out the picture. So when you're showing an ID, it's pretty much fully anonymized, having, like, the picture is missing. The data is covered. It says California driving license. That is not covered and then possibly the question is is this, does this look like a real California driving license, for example? And then the specific question that a human needed to answer can be answered while trying to preserve privacy and security. Then, after this whole process completes, we have to do all the compliance requirements, like you mentioned, ofac and other things.

Yeah, it's packed into this one app that basically follows the guidelines of law and regulations. This is not just the work that we do just for Pi Network. The whole community actually have this produced this large scale KYC to social network. It's actually a contribution from Pi Network from all of us, including pioneers to the world because, as the industry towards getting more and more into the compliance, having a large enough social network that is KYC make the world ready to plug in true Web3 tools or services more easily. It also facilitated more easier onboarding of future other chains or other blockchain or financial services or Web3 type of applications that has a financial aspect or element inside. This is something that why we need to actually build in-house kind of solution, because it's important for our philosophy of accessibility, but also it actually create a great contribution. I mean there was in Web2, facebook's innovation was using real identity, real name in the social network and that break through the last few decades, the fame and the boom of Facebook. The next phase for Web3, web3 means ownership right People are being able to own. When it touches ownership, there is the need of KYC. That's why we're building this huge network, not just large but actually actually more compliant and, towards the future, that easy to plug in for even other services.

That makes a ton of sense to me and I agree with you. Like, if you guys succeed in getting a high quality KYC done, that other people trust other networks, other people who might want to use the like, interact with that network of people, then you've created something pretty valuable for sure. What other kinds of applications are going to run or are already running on PyNetwork? Some chains are purpose-built or better for other things. Then they have unique things that they might be good at, whereas they might not be good at other things. Is it a general purpose, l1, or do you think there are some applications that are more appealing to be built on top of the network than, say, something other applications that might be a better fit somewhere else?

Any type of blockchain application that you know from other networks can also run on Py. There's no restrictions there. We focus also on applications that traditionally have not been popular on blockchains, and I'm talking about applications that are very popular even in Web 2, the way that the world has been working until now. We had a few, a couple of hackathons, and we have some programs inside Py for people to collaborate and create apps, and we've had a few thousand developers engage, producing hundreds of apps and tens of thousands of pioneers giving them feedback on these ideas. So through these they have started building applications of various kinds like, for example, one vertical is remote labor. One example is the KYC app that we build, but another example is apps that are similar to Fiverr or Upwork, where a pioneer can provide some work and in exchange, get paid in Py from someone else, and that work could be from something simple as creating a piece of art to something more complex than this. Another vertical of applications that we see we actually have several marketplaces on Py and these marketplaces each one has a slightly different angle. Some marketplaces are focusing on user-to-user good exchanges, like eBay. Others are focusing on rallying up and connecting merchants around the world, and some are either even creating a decentralized network of logistics. So, because Py is in enclosed domain at period right now and you can't use Py to use FedEx or something like this, people have created this network of individuals who can help with delivery of goods. Other vertical is crowdsourcing tools. So, for example, there is an app for creating polls and someone who wants to get some feedback from community and they don't have a community, they can post a poll there and reward people in Py for answering those questions that they may have. There is other apps such as quote production, funny quotes, good quotes, quality quotes. And there's, of course, games that are using Pi. For example, when you lose, you spend some Pi to continue. Or we use Pi to buy virtual goods, cats and fertilizer or whatever the game has with Pi.

Yeah, so basically overall, the application, in addition to the traditional like blockchain-based kind of application, like DeFi related, which Pi is capable of providing, especially after the open mainnet, we are also emphasizing utility or apps that facilitate access to true goods and services type of apps, and in addition to that utility includes also entertainment or social life, like specifically on the social side applications. There are a bunch also coming from the community developers, but at the same time we also are innovators of social applications and we believe that the true Web3 social apps are not just moving from the Web2 and putting things on chain and then become Web3. Of course that's one aspect, like the ownership aspect, like store things on chain, but we see far beyond that in the sense that we really want to build Web3 applications that actually make improvement of the flaws of Web2, for example, authenticity, the fake news, misinformation, disinformation, for example, information silo, information overload how do you find high signal information from a crowd of noises Like these type of things? We think, with the tool of cryptocurrency and token economics, designing the token economics of using cryptocurrency are able to facilitate the incentive and disincentive inside the social applications to regulate or moderate the behavior online and ultimately achieving better content better curated content and better signal content At the same time, make people more accountable in their behavior. And actually this is not something that we just the only idea. This idea has been with Pi since the last three or four years, but actually recently because our ecosystem reached the point that we're actually about to launch some of our applications. That is an instantiation of such ideas I just described, so we're actually excited. Very soon, some of these apps of ours will actually be released.

And I want to mention that one of the things we can actually do that at Pi is because of the large number of people on the network. So if certain social networking apps, or certain apps that we are describing just now, they only start making sense when you have a large number of people and the utility is much lower when you haven't reached the critical mass.

Yeah, I was going to point that same thing out. You can't have a Fiverr clone on chain without a couple million people, because then you just don't have diversity enough to get the tasks. It's the chicken and egg problem. So these marketplace type applications make a ton of sense. One thing you brought up it was a very quick thing, you said but the thing about information silos, like data silos, this is one of the things that from back in 2017, when they were talking about token curated registries all the time, I was really interested in. There's lots of different data silos. I was just talking to a friend actually yesterday about how he works for a nonprofit that their whole job is to get researchers to open up their databases for collaborative open data, open science, and their issue is it's a social problem. They want to get credit and be the first to produce the final work of the research with that data, and so researchers are just not incentivized to open up their databases. And I agree it'd be so cool to find a token economic approach to solve that problem. Because I'm an open source zealot, by the way, if you didn't know that about me, I love open source technology. It's the only thing I'll ever build is open source software, because I care about the impact that it can have, and I think that you can have the maximum impact with software if it's open source and especially if you weave in decentralized technologies and token economics, all of a sudden there's an economic reason to do it too. That's maybe in theory more sustainable than the way things were before. And I think the other thing I wanted to say too is I liked the way that you framed the transition from Web 2 is just taking an existing thing and making it better, not just completely getting rid of it, and I think that's an important point to make. We're getting towards the top of the show here, and I always ask this last question what have I not asked you guys that you wanted to talk about? What's the thing I should have asked that I didn't ask yet?

Good question.

Do you want to talk about PyOS?

Sure we can talk about PyOS.

So, since you mentioned open source, we recently launched our own open source license agreement. It's called PyOS. This allows developers to get help from other developers and collaborate when they're producing Py applications. We saw a lot of adoption and we saw that it was even helpful in hackathons. So we recently had a hackathon and that resulted in several PyOS applications that were built, and the good thing about it is that in the end of the hackathon, before we start the next one, there is a base, a code base, that, even if half of those applications ended up, end up the developers not continuing, who knows? But the next set of developers can potentially start from that code and make some use of it.

I've seen a few people in our space create their own licenses. I typically go with the AGPL v3 license just mainly because my business partner helped write it back in the day and I like that. It ensures that the things that I work on I'm not an engineer myself, but as an entrepreneur I'm helping get these things built that they'll always be free and open source, like I like that license. I lean very copy left. What is in your license Like? How is it written? What are the benefits of the license? I'd love to read it too. If you can send me the license, I would love to read it. But just like the high level for people, what does your license mean and what are the implications of using it?

It started by prototyping after the MIT license and has some components inside that are emphasizing collaboration within the Pi network.

Yeah, so it facilitates building inside the network with one each other while at the same time protects the network itself as well. So it's a high network specific OS license that's why it's Pi OS and it facilitates within the network contributions to create applications, one on top of the other?

What protections is it providing to the network? What are you guys protecting the network from?

Basically, for example, preventing of just a complete copy of these network's applications to completely just migrate to another network, for example.

It's exclusively within the network this enabled a bunch of developers who otherwise Wouldn't be sharing their code yes to feel more comfortable to actually share the code. We find that conversations when we're giving grants on, when we're discussing with teams of whether they are willing to License their code under pi os we find them much easier to result into. Actually people make yeah it's interesting.

I mean I've had this conversation with so many entrepreneurs and and Developers, you know, who have a lot of fear and anxiety around open source, like just something as simple, like how do I do it? You know, like I want to open source and I don't know what license, but assume I know what license I'm gonna use. How do I do it? And I'm always like you just put it in your repo, like you just make it Open source. It's really easy. There is a lot of philosophical conversation around the different licenses and like the, the protections they provide and Totally get where they have people in your ecosystem who are building these applications are coming from and where you're coming from. My philosophy has always been and I might be wrong I've kind of got this experiment going at this company I'm building, called a olaplex, and I'll tell you guys more about that later. It's not for the show, but it's a hundred percent like full-stacked EPL licensed and my thesis is that by being that way, anybody can fork it and do whatever they want with it, but all the things that they do to improve it will come back and like, if we, if we, can take advantage of those at my company, then we will if that's something that we can use and that really the only way we went as a business is by being open and collaborating with as many people as possible and then Outcompeting them on like BD and sales and marketing right, just getting after it on like acquiring customers, and that as we build those network effects with partnerships and customers, even if you carbon copy my software, it becomes really hard to Compete against us and like even SAS businesses that are closed source. We saw this all through the 2000s when SAS became like a huge thing. Even if you have a closed source software, like someone can come in out, execute you and just copy the whole business, even if they don't have your source code. And so for something like pie I mean you guys have built a massive network is Almost impossible to replicate that part, the network, the software that runs pie could be a hundred percent open source and you just Know. I mean I could copy it, I could figure out how to copy it, but like no one's gonna use it. You know it's like crickets and tumbleweed when I watch it. I could go on about this topic like all day. It's not worth getting like two into the weeds you mentioned like there's some really exciting applications that are launching and I didn't want to leave that one hanging. Like you want to say what some of them are to, like leak some alpha to people who are listening.

Actually is still still so we will, you know, wait, maybe it will be released soon.

Yes, okay, all right, you can. You can leave us hanging, we'll. We will pay attention to Twitter and wait for the announcements. Thanks so much for being on the show, guys. I appreciate it.

Thank you very much. Nice talking to you. Bye, bye.

You just heard the index podcast with your host, alex Kahaya. If you enjoyed this episode, please give the show a five star rating and subscribe on Apple, spotify, google or your favorite streaming platform. New episodes available every other Wednesday. Thanks for tuning in.