Mel Project: Is Web3 Truly Decentralised? - Eric Tung. Ep. 558

Transcript

if we truly had a free market I don't think we would necessarily need all this decentralized technology the market would take care of what we really needed is a new kind of blockchain where what happens on the blockchain can be trustless observed by systems outside the blockchain Bitcoin isn't really made for supporting these offchain use cases so I tried ethereum too uh that didn't turn out that well either Mel started as like this research project to build a blockchain that can actually support a whole internet that's user s in that has like decentralization and security and resilience in all the protocols you know not just the blockchain but if you

actually start building these massive networks that depend on correct knowledge of what happens on the blockchain then if somebody hacks iner for a day that's like worse than hacking ethereum welcome to epicenter the show which talks about the Technologies proes and people driving decentralization and the blockchain revolution I'm Brian Crane and today I'm speaking with Eric Tong who's the founder and CEO and CTO of Mel project and so really excited to get into into that with him there's a lot to unpack but just before we talk with Eric a few words from our sponsor this episode is proudly brought to you by nosis a Visionary Collective

committed to fostering and expanding applications for a decentralized future nosis is at the Forefront of innovation with nosis pay circles and metry revolutionizing open Banking and creating a superior form of money with Hashi and nois VPN they are building a more resilient and privacy focused open internet are you seeking a robust L1 to launch your project well look no further than nosis chain enjoy the same development environment as ethereum but with significantly lower transaction fees and with a robust network of over 200,000 validators nosis chain stands as a incredibly neutral and resilient foundation for your application governance atosis is driven by nosis Dow where everyone has a voice in shaping

the Project's future join the nosis community today by participating in the nosis D governance form you can deploy your project on the evm compatible and highly decentralized nosis chain or help secure the network by running a validator with just a single gno and lowcost Hardware embark on your journey towards decentralization today atosis doio Cars one is one of the biggest node operators globally and help you stake your tokens on 45 plus networks like ethereum Cosmos celesa and dydx more than 100,000 delegators stake with corus one including institutions like bitgo and Ledger sticking with course one not only gets you the highest

years but also the most robust security practices in infrastructure that are usually exclusive for institutions you can St directly to cores one's public note from your wallet set up a white label note or use the recently launched product Opus to stake up to 8,000 eth in a single transaction you can even offer high yield staking to your own customers using their API your assets always remain in your custody so you can have complete Peace of Mind start sing today at chorus. one thanks so much for joining us today it's uh I'm really excited that uh to have you on yeah I'm really excited too so you know we spoke a bit before

and I've looked a bit you know sort of been uh looking a bit into uh Mel or the thing you're working on uh there's a lot there right there's a lot of different components for it it's a very ambitious um it's a very ambitious project but I thought I don't know where it makes sense to start I wonder if it makes sense to start with your VPN and uh or like tell us sort of like your journey of like how you um how he became interested in you know decentralization and the topics around it yeah absolutely so you know I grew up both in China and Canada so so back when I was small my family moved between the

two countries a lot so I kind of got to experience both like a relatively well functioning Western Country but also a country where you know you had a lot of sens censorship a lot of surveillance and all of like technological control over people uh so like that actually uh played a big role in how I got into computers as a whole because when I was small um I was like 11 or so um I was I was just learning how to code and it happened to be the same year that China turned on its great firewall you know and put it into production so you know one day I just wake woke up and then um all of my favorite websites were gone I couldn't go on YouTube I couldn't go on like you know Facebook and um and you

know that was a really big shock to me as a like I need to fix this I needed to go to those websites and I that really got me into things like vpns Anon Anonymous communication technology and really just like there's this whole world that's dedicated to how do you hide from your ISP um how do you uh transmit data on the internet in a way that your ISP doesn't know what's what's going on and so I really kind of dove into that main my I back then with my like very limited coding skills I made like a very janky like browser actually that had a VPN in it had like anti- fishing and things like that into it uh really got me into kind of like um cyber security and privacy and all of that so later on you know I really felt like my

mission in life is really to use cryptography use these tools to build things that let people freely coordinate and freely access information um I didn't know that much about crypto or blockchains I just went to um went to like University at Water Lo later when I was 17 I got into their PHD program I was like very much into like CER Punk things right like I was like very much into tour you know a freet all of that uh how I really got into kind of like decentralization per se is really that I really saw two things so first of all um one thing I really felt like after I started actually researching these things is that you need to build

like trustless systems and this is a you know after like the whole Snowden thing leak happened I realized you couldn't just host your server in the US and expect that to be good enough you can't you you can't just expect okay China's bad um you know Iran's internet's bad but as long as you encrypt your traffic to outside of these countries you're fine well that Paradigm got shattered you you can't trust anybody's infrastructure and the other thing is that and then I discovered Bitcoin which showed that it's actually possible to build things that where you don't trust any particular person through the power of Economics you can actually build systems where you're not assuming that somebody's trusted we even assuming that

a majority is trusted you're just assuming that a majority of people out there want to make money which is a much easier assumption to make and then you can build secure systems so I was super excited about Bitcoin about blockchains I kind of like pivoted my whole Research into how can I build these tools on top of Bitcoin uh I even built like you know public infrastructures and things like that on bitcoin and all this time on the side I was developing my VPN that I started when I was small I rewrote it several times just to I like before you were saying so you started your PhD program at 17 uh yeah that's pretty crazy so you I was H schooled um before um Bachelors before I went to University

So like um I wasn't necessarily limited to a particular schedule so I went to my bachelors when I was 14 and I got into my PhD seven both in the University of waterloop so it was easy I just like stayed at that one School uh and went to whatever Pro liked me the most I guess so and then in the PHD you focused on you were working a lot on uvpn or also other not really like so I was really focusing on PD P networks um in tour and anonymity and that kind of thing and that really led me down to discovering Bitcoin and discovering that you can actually build secure consensus in a decentralized setting and so so yeah

like I kind of pivoted more towards how can we build things like these cool peer-to-peer networks these naming systems on bitcoin but later on I really discovered you kind of can't right like um Bitcoin isn't really made for supporting these offchain use cases so I tried ethereum too uh that didn't turn out that well either and I published a paper or so like here and there but I just felt like that was not going to beow the next generation of the internet so I I really sat and I thought about it I realized what we really needed is a new kind of blockchain where what happens on the blockchain it can be trustless observed

by systems outside the blockchain and that kind of started meel Mel started as like this research project to build a blockchain back at actually support a whole internet that's user Sovereign that's you know um that has like decentralization and security and resilience in all the protocols you know not just the blockchain but you need a blockchain to support that sister so you said you were kind of like trying out ethereum and then you felt like ethereum wasn't the answer what was the what were the things that you felt were the shortcomings of ethereum there's only just honestly and it's not just ethereum it's every other layer one

which is that we don't have really good light clients so here's an example so let's say I'm building a bit torrent replacement except um the nodes are incentivized so you know if I download a file from I got to pay you something like that let's say you want to build this on ethereum well there's many ways you can do this you can just use onchain transfers you could use some cool state state Channel thing cool Layer Two it doesn't matter there's lots of ways you can do this but if but if you actually get down to implementing this so you have a born client running on your computer and it's downloading a file from a peer and it needs to pay that pier and it also needs to verify incoming payments so all of these things

go through an RPC provider like infura um now this has two problems first of all they see everything you do they can censor it but that's not even the worst problem the worst problem is that they can lie to you right like when you asking did I get my payment they can say yes when the actual answer is no and there's no way for your client to know so really the whole security the whole economics of the system completely depends on etheria telling you truthfully what's happening on an ethereum blockchain and then your system's no longer decentralized I thought the torrent was supposed to be decentralized now you add a blockchain in and now it's centralized this that's the opposite of what you're trying to do right like so so like I was trying to

build these systems in with every single L Tread Like This became a huge problem and I suspect that the reason why we currently don't feel like this is a huge problem is not is because we're not building this kind of system that actually is offchain but uses onchain for security because if you're Building Systems that are on chain then this is fine right if you're transferring entities on chain if if you're allies to you they don't steal any money they don't do any damage you go switch to a different RPC provider if somebody hacks in for a day it'll be a great inconvenience it wouldn't lead to massive hacks of huge worldwide systems but if you actually start building these

massive networks that depend on correct knowledge of what happens on the blockchain then if somebody hacks infer for a day that's like worse than hacking ethereum of course I think a lot of people are sort of like aware of this problem I mean I remember a great article by probably put the link on that in the show notes I think we mentioned it before but there's this great article by like Moxy M Marin I'm sure exactly yeah right but he kind of points exactly to this issue but yeah you're right that like people don't tend to think of this as a problem because I guess why would I mean if you're lying to you it would ruin their whole business they don't have an incentive to

do that right but I guess I wonder if if the would the issue be more something like okay let's say if you know build this Bitcoin network uh that has ethereum incentives and you depending on a centralized RBC for Rider I mean I imagine where the challenge might come is that they're going to be like oh but you know we don't want to you know from a regulatory perspective we don't want to like support this kind of thing so we going to have to like thr this or shut it down or yeah I mean honestly this is the whole point of crypto right I mean if you think about it it's not in the bank's incentive to close your account it's not in the New York Stock Exchange

it's incentive to stop you from selling dogecoins on it like the whole idea is like if we like if we truly had a free market I don't think we would necessarily need all this decentralized technology the market would take care of it the problem is that we don't right we have monopolies and oligopolies that are control that that you know have a symbiotic relationship with regul regulators and that's the whole problem we're solving with crypto so you know so basically here's the thing like if you're building a app that depends on AA you know if you if that's fine for your use case then it's also fine not to use ethereum it's also fine to have inferior run the whole app

infrastructure for you right like um now this is different so I think that the there's the biggest reason why people don't feel like this is a problem is because we're not trying to build this kind of app where the security bottleneck actually is in furo um so for example if we're actually doing you know defy trating where like the current web 3 ecosystem infura isn't really the security bottleneck because as I alluded to if infura lies to you about how many pudy penguins you have how many e you have what's the status of the pools and gunis swop the Damage they can do is very limited right like why would they do so and ruin their own reputation uh the worst thing that could happen maybe is that they censor your transaction

they're not because of regulatory reasons they're not going to lie to you because it's not in their interest and not in the interest of anybody who hacks inferia like like there's no reason why I would hack inferia to troll people to say everybody has a pudgy penguin like maybe I would want to do that for fun it's not going to make me a ton of money yeah and and I guess the other thing I could also Imagine happening here now in the example of like okay bit torn black network is that of course some people not like that right like let's say whole copyright type stuff right so I could imagine then them basically being like hey we going to try to force INF fura to give the IP addresses right of the people so that they can basically go and uh P them yeah

and I think like that's definitely a big problem but I actually think that that's not the worst problem here because that's still like a sensorship problem of inur stopping the network I think the bigger problem is lying so I think maybe bit torn payment verification is not necessarily the best example uh so there have other use cases like for example into an encryption so I need to look up Brian crane like like What mo Moy Marlin bike talks about right I look up your public key on a blockchain and I use that to do into an encryption now if infer lies to me they can steal every single message I sent to you right like maybe I'm doing some like big business deal here with and transferring a bunch

of money like you can literally steal money this way um by lying to the users and so so and it really boils down to when we're using these decentralized systems we expect that we're trusting the blockchain's economics we're not expecting that we're simply running we're simply using a system run by insur so like and you know because of this if you're trying to build this whole internet where where the security really comes from the blockchain and you're building these offchain systems like you know payment networks or you know B torent light networks and n and encrypted Messengers then knowing what happens on chain really

matters and the whole thesis of metal is that these use cases are way bigger than the current web three ecosystem the current web three ecosystem is like let's stuff things on chain do the composability onchain because that's the safe place right that's where all the security consensus happens once you're outside of it you're kind kind of like uh out of luck but if you can't actually transfer the security out of it then this changes the whole Paradigm it makes building out web 3 and mass adoption everything way easier because you just need to look at the blockchain from outside when when you need to so I I really like what you said before like you know if you had it truly free market

we wouldn't need uh all this decentralization and I I think that's a very true point and I think on some level right if you kind of zoom out right like what is all this crypto blockchain decentral decentralization about I to me is is also about creating uh an economic system that is actually a free market exactly and I think that like Mel is really about about that as well I think what crypto has currently done is it has greatly liberated Finance right anyone can create Financial instrument create tokens sell basically stock in projects right um but the but I think that the

rest of the internet needs a similar change you want a whole internet that's permissionless that's borderless that people can transact and communicate on freely and what Mel does is that it takes all the cool security and decentralization we've invented for blockchains and uses like cines to let people build other protocols like communication protocols like public protocols to have the same kind of properties and really get get the web Revolution happening across the internet and Crea and this is like making a truly free market right because you don't you can't just only have a free way of sending money you also need free communication you know free publishing all of that so that people can

coordinate and um come to consensus on things I I would love it actually if you can go a bit more into dep on you know the kind of like use cases that you feel like you know cuz you're saying okay this is the kind of thing that you know could really get us to like Mass adoption and to much broader use cases than the current sort of you know defi type thing like what are the use cases that you feel you know those are really the things that are most powerful that could be enabled by such a technology and that would really you know be things that a lot of people would start uh using that you know

today not using blockchain or maybe not using blockchain for these particular types of problems yeah definitely so I think that it would be helpful to start with the actual use kits right so imagine you had a Discord like app that's completely censorship resistant and permissionless so you know it's the U imagine the ux is the same it's just Discord except you know you don't have to trust the Discord company while using it you can say whatever you want on it start your own communities on it and you know like you know and basically just like create this whole create all these communities on this system that are

sovereign and our user focused so let's just imagine an app like that so the the societal impact of this would be really big on because if we think about currently how can you get people to work together there are basically two ways one is kind of like a in a private informal volunteer basis way so I I'm a person living in the US you're living in China he's living in Russia let's go on GitHub and try to write some code together try to build something and this this is one model the other model is that this fully regulated very bureaucratic multinational corporation where you know I have to like you know make this whole legal structure uh hire people do some some big thing together

and it's all very complex and all very like it's really easily um damaged or shut down by like random like political or like societal differences so that's the current world but if you have a truly permissionless communication and Community Building tool what you can actually build as an example would be a multinational borderless permissionless Collective of people working on things that's properly governed and properly incentivized so imagine you can do votes on this I mean Discord can do votes you can do votes on this you can use this to control the treasury talk about things completely permissionless and you can

access this app anywhere in the world you know China can't block it Iran can't block it uh nobody can stop your money from flowing nobody can see what you're doing on it so this will do to community building and DOW creation and like coordination what you know Bitcoin did to money transfer so you could imagine people working on things like let's let's archive all the books or let's you know uh share um whistleblower information or or do journalism in these communities and it'll be both public but also permissionless and organic while currently we have a dichotomy between if you want secure and permissionless and kind of like outside the it cannot be

shut down it has to be low profile and private if you want public then that's very like bureaucratic and regulated if you have a it's just this one application will let you build a world where people can make public but also permissionless collectives that work on things together yeah so the problem with something like Discord today is like okay if you I I mean I guess so I guess one problem would be how do you link the Discord to something now you know uh uh onchain like you know Rec control a bunch of funds together things like that right so so that that's not the

biggest problem I think the biggest problem is that Discord is a centralized platform right like Discord sees everything you do they can shut down a community they don't like and you know also like if your ISP doesn't like you accessing Discord then you don't catch to access Discord like in China it's quite difficult to access Discord So like um I think like it's not really about the onchain it's really about the software itself you know what can you do if in a in web 2 versus a truly permissionless application and so so if you think about this permissionless Discord just like the actual Discord parts let's forget about crypto or blockchains I think that'll really create this very vibrant Marketplace of ideas of coordination that currently

doesn't exist on the internet we can really let people from all around the world join and participate and do things we can't currently just can't do and I think that what and and the thing about Mel is that to realize this kind of vision you need two things first of all you need a way of actually combining these centralization but also like a centralized user experience right like the problem with decentralized Messengers now is that you got to like manually set up your servers and things like that and you need to manually connect to them and you're still trusting that particular server um but with a with like clients in a blockchain like Mel you can actually

have a whole free market of different providers you can pick from move your group between these things um and I wrote a whole blog post on on this called like confederal protocols where I explore this concept where it is that we start with a model that we know how to build with is which is Federated message passing which was used in like email used in Matrix used in xmpp and we simply had a very simp simple blockchain layer to that and we get like fully user Sovereign very decentralized um in better user experience because the whole thing becomes one name space it's like Discord you can search for servers join them you have one username you have like one identity your identity is not bound

to whatever server you pick and so you end up having the same user experience as a centralized app except it's permissionless all the incentives actually go to your community not to Discord so imagine when you boost your server you're actually contributing to public goods for your community and it's sensorship resistant and like it it's going to have better quality of service because there's going to be a pure free market in the service providers that provide hosting for these communities so and because switching is basically zero cost and so that's kind of thing that Mel lets you build like this kind of like much fre your market for resources for protocol infrastructure that let you

build much better internet apps than with web 2 or I guess with current web 3 technology so maybe we can go a bit into a little bit into the light client aspect here because I mean light clients is something that of course in crypto it's always you know everyone's always or you know people understand like kind of the the importance of like clients at at least more or less I think uh what is why is it so hard to have for existing blockchains to have good light clients and like what's the you know

what's the challenge around light clients and what how do you do light clients and and how come they're so much more efficient and and maybe also kind of a a connected question I mean to me it seems like this is something where actually there's a lot of uh innovational work at the moment that that goes into this kind of thing of like hey have really efficient proofs of what happens onchain offchain which is I guess like ZK proofs uh do you feel like ZK proofs are kind of enable this properties for like any blockchain or is this yeah yeah so maybe like talk a little bit about like the kind of the

role of light clients and how the technical challenges around efficient light clients yeah so I think that like the thing about Li clients is that it is true that let's say if you're wanting to build a li client for ethereum there's a ton of technical challenges you have to overcome because ethereum is a complex protocol you know you need to somehow prove that somebody correctly executed all of that and show it to your like client that is actually very difficult so how Mel so the thing about Mel is that it's not really dealing with the same problems as like clients for existing l1s are dealing with um the the the thing about Mel is that because we're making a new layer one we can

design everything in a way that it's easy to prove to like clients so you know there's a ton of things that goes into it like putting everything in Merle trees you know using a very simple utxo model um the upshot is that it's it actually turns out to be very easy to prove to your live client that we came to consensus on this this part of the state says this without using ZK proofs now of course ZK proofs can compress this but even without that it's very efficient um and I think that so so this is one part like how easy it is to actually prove the states the other part is actually like I think that there's a lot of awareness in this space that like

clients are good but I think that what LEL really sees is why like what can you do with lines and I think that currently in this blockchain space we tend to think of like clients as a scaling technique for um you know like doing things like crash chain things it's really about how do you make the blockchain better but M's Insight is that if you have really really good like so you just it's not just enough to have light light clients you need the light clients to for example be very very light and you know allow them to for example go offline for long periods of time without losing security you combine a lot of these features you have this

like super powered like client then you can build categorically different and much bigger and better applications than what the blockchain ecosystem is focused on right now so if you think about kind of like this sort of decentralized Discord let's just think about the architecture of this it's actually very simple you simply start with a system like Matrix and you put a mapping between your username and which server you're on onto the blockchain so you know for example currently May Brian crane is at matrix.org tomorrow it might be at Brian crane.com you know this mapping on chain and in your actual like chat client you look up

this mapping to see where which server to contact to talk to this guy so it's literally just this very simple little thing on the blockchain and you you have to be able to look at that securely with lines but after that it's a completely conventional traditional internet system it's just a Federated chat system just like Matrix or email and so Mel's Insight is that if you have good light clients you can start building things like this and you can easily build full stack decentralized user Sovereign software by putting a little bit of things on the blockchain and having every single user look at it and I think like I think that's something that's

hard even if you have let's say ethereum light CL because ethereum like clients for example they have to be online all the time they're not that light um you know other blockchains have similar caveat so if you're like client they they T even when they do have like clients they tend to not be optimized for this kind of use case because that's not what they're thinking about when they're designing the like so that that means you can't build a system like this practically um on these other blockchains but you can on know and I think this is the key ins section so in this uh example of the the Discord like thing so the thing that you would uh put on chain so you said a mapping of usernames so let's say you

know cran VF or some sort of thing like that you you said the mapping of usernames and servers you put on can you explain that a little bit more yeah definitely so um so let's just use email as an example right so currently you might have crane BF example.com and that that's your identity so your identity is owned by example.com and this means that you don't have a free market in providers if you switch to a different provider then you don't keep your identity so this is the first problem so the first problem is that you you don't own your identity somebody else owns your identity the other problem is that um endtoend encryption is hard you might

use pgp to encrypt your emails but how are how are people going to know your public hey right so um so what ends up happening is that you don't have good encryption and you don't control your identity and basically the the situations that the provider has sovereignty over they see everything they own you you don't own your own yourself the like example.com owns you now let's just add to this email system a very basic layer which is a mapping between crane BF like a global username not cran BF at example.com so cran BF with an ens like system map map that to example comma Republic so you put that on chain so so

so here's here's what'll happen then now when I want to talk to cran BF I just need to make send a message to cran BF and my chat client immediately knows which server to contact and what public he to encrypt my message with so I meant send the message UK cran BF and you get it and you know I don't have to know your server or your public here and I don't have to trust anybody else to tell it it's trustless stored on the blockchain and I looked at it with my like client so I verified the information myself now tomorrow let's say example.com um becomes this Draconian censorship machine they don't like you anymore so you switch great I'm going to go to like fub bar. XYZ so all

I need to do is register an account at Fubar cran BF at Fubar XYZ point my um thing to f. XYZ and continue using the software I don't have to like my identity is that is not owned by example.com I'm just like they have to serve me I'm just their customer I can go switch to anybody else so what this really means is that there's going to be like free market in providers you can switch with no cost and the providers don't can't decrypt your messages all they're doing is offering storage and relaying services and if you don't like the server you can switch to somebody else and so so you have so architecturally this is still email it's

literally just email but with this little anching bit that the clients look at and but because of this then this whole system is user Sovereign end to encrypted and honestly decentralized in trusts it's web 3 it's no longer web 2 by the addition of this little piece of piece on the blockchain so I think this is really what Mel is about you can start building these systems where you can sprinkle this like web 3 pixie dust on on a traditional web one protocol and you get something that's radically more secure more user Sovereign more decentralized and this only works if the the the kind of like offchain Parts can compose with the onchain parts securely using

lines so we talked a bit about this um thing of like basically the communication and the sort of community aspect what what are the other use cases that you're most excited about about this technology enabling well I think that the thing is that uh it's going to compose like layer by layer right like if you think about how do humans interact there's really two big things one is transactions I need to be able to send you money the other is communication I need to talk to you now if we can talk to each other and send each other money we can do business and create value and create things so what Mel is really about is building the

infrastructure to allow this kind of value creation that unlock these opportunities of people interacting and Building Things That previously don't exist so you know you could imagine some G on this Discord like thing comes up with some really cool idea some public good funding system let's say and they they're operating that well that's a use case that's only enabled by Mel but it's not directly it's more like we have we build an infrastructure to let people interact and that's just like the how the internet works right the internet doesn't it's not really like Riverside FM is a ipv4 use case well technically it is but it's not like ipv4 is designed so that okay one use case would be

riverside. FM it's more like it's such a general purpose liberating tool that allows people to innovate systems like Riverside out F and you know of course you can think of specific things you can build on M other than the Discord thing for example I've also written blog post sketching out how you can build decentralized incentivized MMO games on uh in this Paradigm how you can you can also build things like you know much better file storage networks that are truly free market and not based on tokenomics unlike you know file Queen Etc but truly based on you just pay your provider these are all things you can easily build within this whole Paradigm and I think that we already know how to

innovate like this because it's just the internet it's the whole internet Paradigm the only difference is that now with like clients you can get decentralized consensus on secur security critical things whenever you need it but once you have that you don't need to reinvent the wheel we know how to build federat protocols we know how to build peer-to-peer file transfer protocols we know how to build web hosting protocols we know how to host HTTP servers we just need to compose those with pieces that let us incentivize and decentralize all you know one of the main things that people um do on blockchains is you know smart contracts and you know complete smart

contracts and then build all kinds of stuff on that is that something that male the mail chain is also able to do or like what's your view on like smart contracts well I think they're excellent abstraction for what they're used for right now which is Defi and making financial instruments um I mean Mel technically has two incomplete smart contracts even though it has a utxl model so you could do it on layer one but I would imagine in the Mel ecosystem um most people would not be doing this on layer one um it would be let's say on a ZK roll up on Mel that is evm compatible so that kind of thing would work much better um so that's how I think about it I think it's a great tool

as its job I don't think it should be the world computer and it shouldn't be what defines web 3 it's simply one kind of decentralized app that Mak sense in the ecosystem yeah and it shouldn't be the world computer because of like um one is it like scalability issues if you try to put all the smart contracts on the Chain uh or it's also because you feel like just by having light clients you can do a lot more things that are you know offchain where today you do it onchain like the whole reason why we feel like scalability is a problem is because we have to do everything on CH so all we got to do we got to somehow fit it in right um Mel

whole thing and that's hard right like even if you have a very very very fast blockchain you still need to convince the whole world to this new paradigm right don't run servers don't run peer-to-peer nodes stuff all your logic on chain and trust that the nodes on whatever bloing will run it for you I mean that's like how ICP was trying to do it for example um that's difficult like that I don't think that's like even if it could work it's very difficult it requires rewriting the whole internet Mel's approach is actually that you know given that we have good like clients why do we need to do that we can build our internet protocols the way people have done for decades we know how to do that in a scalable and sustainable way except

that now we can make those things decentralized and secure whenever we need to by reaching for the blockchain to compose with onchain logic whenever we need to but but not by using the blockchain as the world computer that the whole ecosystem lives on now that's a paradigm that I don't think generally works and has worked in the past one analogy I like to use is the internet versus Telecom before the internet so Telecom before the internet is basically this one world phone network approach the whole world was one integrated Telecom system from application to the infrastructure um all the Innovation happened on the Bas layer and you know you want better quality phone calls

baselayer you want you know pagers so that your uh a device can buzz when you get a phone call that's the job for the Bas layer everything requires infrastructure level Bas layer improvements now what the internet really realize is that actually we can make the bay layer composable simple and neutral it's just a packet routing layer you know it doesn't need to have any features it just needs to sit there and move stuff and you know compose with other networks so it's actually decentralized in way the internet and then we can leave as long we can allow anything to use it across the Telecom stock whenever they need to and then we're done right like Market Innovation takes care of the rest AT&T did not need

to invent every single website and I think we need we're going to see something very similar in web 3 which is that the early Paradigm of doing everything in the Bas layer that's easy to build out at first but if really is inherently limiting because it means that all sorts of key Innovations require improving the Bas layer like you know adding CK stuff improving throughput um you know improving storage capacity and it's actually worse with blockchains because every time you change the base layer it's governance and you run into the issue of if you change a base layer all the time is it even decentralized or is it you know actually controlled by the de team uh but even ignoring that right this is just architecturally like very slow and

very like limiting if we can have like a stable base layer that just sits there so that you can put things on it that need Global consensus and you innovate and build protocols elsewhere I think that's copying the success story of the internet um that really took Telecom to like unimaginable new level by decoupling the Bas layer from the apps so I think like that's how um I'm thinking about where M's going and then so it said you know the thing mail that you're building here it's a proof of stake uh layer one so let's talk about this a little bit so it's utxo uh proof for sake base

layer one so for example for consensus how does the consensus work well we use like a Byzantine fall toering consensus protocol called streamlet so I as far as I'm aware we might be the only blockchain to actually use that uh in a production system so streamlet is a bft algorithm invented in 2020 that's actually designed for pedagogy like you know teaching people how bft works so it's a bft protocol designed to be as simple and easy to explain as possible and we chose that because we wanted we want our base layer to be as simple and easy to explain and easy to implement as possible uh so we picked this kind of bft algorithm um I really suggest

everyone read the paper like of Streamlight it's surpris it's shocking how simple you can make a bft algorithm like it's basically as simple as bitcoin's consensus except it's proof of stake and it achieves finality um so we need finality for good my clients so we we pet stream yeah and and then issue I mean a lot of people will be familiar with tender Min you know which is I think sort of like the generally also considered like you know a very simple um bft proof for sake system so but streamlet is even simpler or like in what ways is it simpler than something like tenament well uh I think it's kind

of hard to explain it is just literally way simpler like the paper is much shorter it's much more obvious after you read it how to implement it and why it's correct um so it's actually built on top of a lot of previous work including tendermint um it's really the product of a lot of work done by the office of that paper to really build the easiest to explain and Implement bft algorithm um so yeah like with tenderman for example there's these several rounds of voting and all of that you know it's kind of tricky to understand why that's that you need to do that and how to implement it correctly with streamlet it's much easier so um it's also in lineage I would say it's closer to like hot stuff

and uh Casper bft it's more like uh kind of like a chain based bft rather than a voting based bft So like um but but that's like more of a technical detail but I like I think that it's by far like the simplest bft algorithm I've seen and after I implemented I realized its performance characteristics were good enough for our use case so we decided to use them okay so yeah so we have this simple proof of stake bft system we also maybe we can talk about the utxo thing briefly uh I mean utxo of course you know everyone knows like Bitcoin uses utxo there are other chains right that

have tried to do utxo things for smart contracts but like what are the pros and cons between the sort of account based model that ethereum has and utxo based model well how I'm thinking about it is that if you're trying to do smart contracts if you're trying to make an ethereum killer don't use the utxos it's a bad idea like that's you know like I don't think utx is are the right model if you're trying to do a basically an account-based smart contract model like it's trying to simulate that on utxo is possible but don't try that it's not a good idea uh there are other BLS that have tried that and I don't think it's a great idea uh now why did M use cxos is

precisely because it's trying to do something very different from blockchains like etheria um the the big advantage of the utxl model that there the reason why Mel uses it is that it makes all the state and all the data on chain structured rather than unstructured so if you think about let's say an ethereum if you're building a system like ens um even if you have like clients how do you know what name you know creting the fs map to you have to call a contract and run evm code and that evm code accesses something in the state dynamically and tells you the answer

that's just inherently very complicated if you have a like client that means many many round trips to request little pieces of data your contract needs to execute it means implementing a whole evm interpreter in your like client and all of this is because the ontain data in a smart contract based mod like account based model is leg to the onchain code it's not legible to somebody just looking at the blockchain from the outside with the utxl model that's very different because the actual utoo graph is structured you can Traverse it objectively without running onchain code so you can design your contract to Simply encode let's say your

mapping in the structure of the utxo graph and then we can just look up that structure Traverse it with a like client without running onchain code without diving into little pieces onchain so that's the number one reason why we use the utxl model I think it's the best model if your main use case is to expose data in a way legible to offchain by CS because you can basically I I guess if I can paraphrase it if I get it correctly right in in ethereum basically because you don't really you know you have different transactions and you don't know what they like you know you sort of have to know the state of the chain to know how

that transaction affects like other parts of the chain yeah exactly it's not explicitly spelled out yeah yeah whereas here you could just basically take the utxos that you know relate to a particular uh username and then as long as you see those utxos then like you can kind of know yeah and you can also Trace their history back you see see where they came from and you can use like onchain contracts to encode a shape it like for example I actually once wrote a paper a long time ago of encoding a naming system um in the utxo graph of Bitcoin so you literally can use the utxo graph as like a binary search tree and like build a tree and then you can you know you can s like what 0 one0 one0

one that means Brian whatever and then that's where your binding is so you can do tricks like this because the utx graph is structured data not unstructured data I mean I guess one thing that it could be interesting to go into but I don't know if if that's where we should go now uh I know you also have this sort of low volatility asset but I don't know if it should go there or if if there's some other parts of the system that you think are more important to explain now well I I think that like um I mean that that that might be interesting but I feel like I feel like the most interesting thing about Mel is not actually kind of its individual technical parts right because the thing about Mel is that I think it's

really easy to like misunderstand it a little bit as okay it has all these cool Parts there is a lot of cool Parts it has a different consensus a different uh currency Etc but I think what's really cool about Mal is What it lets you build right so that's why for example right now we're actually focusing mostly on building arendal which is this offchain data and Money Transfer layer that is built on Mel and that'll really be the like protocols like a rendle are actually where the ecosystem is going to be built around not like apps building on layer one or like people running notes on layer one so I think like that's an interesting thing about M it has a lot of cool Parts but I think like

what explains the design decisions I make with all these parts is this overall idea that how do you make a blockchain that you can use to build systems like you know a rendol like you know the Discord thing I mentioned earlier so yeah so let's talk about a Randle what's air Randle well so air rendle is like if you think about okay so the I think air rendle could be explained as like this way which is like it's a network that you can use to transfer money and data both quickly and privately so you can think of it as kind of like tour plus lightning Network so you know you can plug your app app into it and send data across or send money across and it's all

offchain and fast uh there but then there's two big differences between air rendle and t plus lightning Network the first one is that the nodes are incentivized and the nodes are incentivized not through like a deep and like token system it's actually very simple which is that whenever you use resources on a node you pay them you pay them with this same payment Channel system so for example I send a pack set a DAT packet of data through you your node and I have to pay a price to use your node resources and and we actually negotiate this price with each other so it's a free market there's no like protocol that coordinates all these prices people find peers that they're

willing to peer with at prices they are willing to accept and the second pillar is that air Brendle is the first peer-to-peer Network designed to be censorship resistant against isps and this is actually really unprecedented um because we think about censorship resistance a lot in web 3 but we typically mean the other nodes on the network cannot sensor right like Monero nodes can't pick and choose who who gets through we don't mean the ISP can't stop him from using the protocol he can ban the whole protocol and this this kind of sensorship resistant I call ban resistance which is like it's resistant to isp's Banning the protocol and it's

very difficult to do that in a decentralized setting but I think it's very important because you know we're we're trying to build permissionless assistance here right we're not we're not saying okay we just don't we we don't trust our bank but we trust the ISP that's not what we're trying to do we're trying to build permissionless systems that and that means we need to be sensorship resistant against our ISP as well and so a rendle uses a lot of Technology I've developed for GE you know my VPN project a lot of the research I've done in traffic confiscation in sensorship resistance to build the world's first decentralized ISP censorship resistant um transportation Network so that's really what air rendle is is at a high level um

and that's only possible because we can build on top of smell like you know the micro payments go on payment channels kind of like lightning Network um but settling on Mel and because Mel has much better light clients we can do so without forcing people to run heavyweight infrastructure know with lightning networks if you run lightning node it's basically a Bitcoin full node it's very difficult to do so um on your phone for example uh but with Mel you can you build payment Channel networks where the clients are very lightweight and can really fit in every single endpoint so that's just like one example of how we use Mel but basically the incentive system and the kind of like security

systems they're all built on top of M CS what's the state of the project now like what what have you buil and like where are you sort of in in the road map to making this happen well like currently um for a rendle we're currently in an alphanet kind of situation we have like a few dozen nodes running a lot of them run by community members uh and overall like we're that's the current focus of our development so we're quickly iterating upon of it upon it we plan to launch the air renal Network into production later this year and then launching air rendle into production that means but that's that

comes before because the male proof of stake chain is not in production yet well the male proof of steak has a Dev net right and that has in public for several years so the initial version of Mir rendle would run on the Mel devet but eventually we're going to upgrade that Dev net into a main net so like the dev nut wouldn't be shut down it would just be relabeled as the main nut once it's ready and what's the what's the sort of difference between the devet and the point when it becomes a main net well I think the biggest uh um difference is that right now we rent all the validators we nobody else has any tokens right like um so that's the main difference once we actually decentralize the token ownership and allow other

people to lend validators and decentralize that then it becomes the main yeah okay okay and then what what about like other parties building applications on Mel like what's the state of that are there people building things and like is it sort of is there a developer ecosystem already so we have a whole SDK for building things ONN but kind of in the overall road map of the project right now is it we're not not quite there yet with like getting people to build an ecosystem and the reason for this is because you know if you think about a

typical crypto layer one launching the devet is like closer to the last step right it's ready to develop um with Mel that's not the case because the ecosystem is not going to be built by people just running things on layer one it's really about people composing with lowlevel protocols within the ecosystem like a rendle so I think like um the how we're going to do it is that we're going to first a rle with a big use case for a rendle which is actually gu my VPN project will turn into a frontend for a rendle so I actually move all of my users from g into air the air rendle ecosystem by just upgrading their software because

the user experience is not going to change they're just going to be using a rendle and that's like about you know half a million or more daily active users they're going to be using this first big use case of a rendle and then I think like developer community commities and people building on the system they would naturally look to compose with this ecosystem all the revenue generation happening on air rle building new applications on air rendle building other protocols like air rendle using it as an example so I think that that's the phase in which it makes sense to start um having applications built on this ecosystem not when there's only the layer one um this is kind of like the internet right like if you only have ipv4 it's not really ready to build

things on but once we had email once we had tcpip yeah that's when the developing ecosystem started form so I mean you mentioned the the VPN right called G so a Randle basically would would be able to perform all of the the kind of functions that people currently use the VPN for yes and actually a lot more than that because with vpns you can only access websites right uh with a rle you're Al also able to host websites anonymously and also browse other websites that people host on the a rle network and also send money to each other so it's like a general purpose communication and payment

Network and yeah it's like a suep of what bpn are able to do okay so so you can host website anonymously of course uh you know we were like aware of the example of that is like tour today right where you can yes yes it's quite similar to the interface of tour yeah and then is it also like you know I think in your tour website you know like you're like the hosting websites anonymously they also wouldn't have you know normal it's not going to be like ww. like whatever but it's going to have its own domain names yes um so in tour for example you have this long string of characters.

onion that encodes your public heat and now the cool thing is that because we have M light clients we don't need to do that we can just be a short human readable name. Haven which is what air is because you know we can store a naming system like as I as mentioned that binds your name to your public key so we can actually have really human readable domain names and website names in this ecosystem I mean I guess this is is also a potential use case or interesting use case would be kind of you know replacing or competing with tour yeah absolutely I think that like um ambitiously I feel like a rendle is

going to kill tour and the reason is because like it's more decentralized um because there's no like Central directory server it's also way better incentivized because if you actually serve the network you get the market rate of like what people are willing to pay you the problem with tour is that it's all volunteer run which means that the only people running it are basically people with way too much money which are basically spy agencies like because why would anybody else actually do this for free it's a lot of B with a lot of risk um so like the I think it'll it'll definitely have better performance than tour because it's also not going to be always overloaded because of this incentive system and also the the

application scope is much bigger than T because T is really just focused on one particular kind of anonymity um while a rendle is actually designed so that the user can select how much anonymity they want they can get toy like anonymity they can get much better like Nim for example they can use mixet delays to get much better anonymity they can also get very low anonymity for high performance they can just use this as a general purpose internet and only for ISP sensorship resistance not for anonymity and you can build like very high performance systems on it as fast as like you know any other system so I think like um it's really what air Bendle is is that

it's going to be this Universal internet layer that has the same interface as the Internet it's like a virtual Network that you can run any application you want on so yeah I think like that's I guess what a rendle is it's really doing it's kind of like combining what people already do with tour with Nim with a regular internet with other P Li P2P all of these protocols uh in this one general purpose elegant abstraction cool cool well let's maybe let's talk a little bit about um your low volatility kind of stable coin why create something like that well

um because I want cryptocurrency that's the short answer I me if you think about it right like why did people inv Bitcoin they wanted a nonfat currency like a unit of value a like um medium of exchange hisor like and all of that right so the thing is that we try with Bitcoin we tried with other crypto it didn't work every single one of them turned into a uh speculation token it's essentially a mem coin in dis disguise nobody uses Bitcoin as the way they use dollars right so like and the reason for that is because they all have fixed Supply and they and you know people's demand is fall

so the price has to be volatile so what Mel is trying to build is I want to build an actual cryptocurrency I want to build something that's trustless that's decentralized but also feels like a currency it's price shouldn't be too crazy go up and down people should be mostly using it to pay for things not speculate on like it's going to go to the moon right um and that's the reason why La melt the currency uh and the objective here is that I I don't want to I don't need to make a stable coin I just need to make a coin that's stable enough that it has a relatively St but you can do Accounting in it you can price things

reasonably so you want the price to look something like US dollar to Euro rather than US dollar to bitcoin and you also can't use oracles because that'll make it non decentralized and I realized there's nothing else in the ecosystem like this so I sat of made made my own design and what's that design like yeah so the high level objective is that I want to Peg one m to this unit called the dok the day of sequential computation so a day of sequential computation is the cost of renting the fastest processor core you can find and occupying it for a day so that's what a

dok is so the interesting thing about this metric is St it's one of the only metric St have these two properties first of all it's actually stable in purchasing power over time so 20 years ago renting the fastest computer back then for a day cost about the same amount of money as right now renting the fastest computer right now for a day in short the fastest processor out there is about the same price to rent at any given point in time so it's stable the second thing is that it's Stressless measurable and that's the really cool part which is that we do not need an oracle to tell us how much money this is and this is because what we can do is

that we can design an auction on chain where we give away some tokens and we say give me proofs of sequential work so you can actually compute like essentially a zero knowledge proof that you did a bunch of nested hashes which has to be done sequentially and then you can prove okay I did this many hashes and this many hours give me the reward so what this really gives you is two things first of all you get an exchange rate between that token and how many hashes people are willing to give up to get that token this is Market base right you don't need the Oracle for this you can basically you can get an exchange rate between

token and hashes and then by how fast people are doing these proofs you can get get an exchange rate between hashes and time so finally if you combine this you that get an exchange rate between your token and time and then you can use this to drive a a stable coin mechanism so that roughly one token equals 24 hours does that make sense yeah but does that mean is it does it mean it's basically kind of like a proof of work asset still because like to get it you you do the you do this sequential computation and then you get a token so if I want want to get a lot of tokens I'm going to run a lot of these CPUs each for a day it's

it's a little like it's proof of work in that sense right like you use proof of work to Mint these tokens but it it's not where the consensus comes from and I also don't expect there to be lots of people doing this the reason is because in in typical proof of work the way it works is that you're printing these tokens and um people kind of like divide up a fixed Supply by how much like proof of work they do and you if the token goes up then like people waste more and more power like competing over this same pole of tokens with M's approach is more like if you do this much computation you're guaranteed this much tokens um so what'll happen is that this will increase the supply of the tokens until

it's no longer profitable to Mint and then people will stop and so the minting is only used as an Arbitrage mechanism to keep the peg in check it's not going to be profitable most of the time it's only profitable when the protocol needs you to Min so that you push down the price but you still going to have to case that like I don't know so even if it's like roughly stable right like it's someone else and maybe it cost them you know $1 to produce like one token someone else it cost them like you know 90 cents or 80 cents or something like that so 80 guy is going to Mint the $1 guy is not going to Mint right right right and then 80 cents guy is going to

mint mint mint mint mint until there's so many tokens that the tokens worth 80 cents and then he's going to I mean I guess you still have kind of the downside here that you know if now male gets a lot of adoption a lot of usage then it still means that a lot of basically a similar amount of I mean let's say now the M tone Supply needs to get to like you know five billion because of how much usage there is of the thing then it still means that there's basically roughly $5 billion worth of money will be spent or you know 4 and a half or whatever uh to get those to get all those tokens in Supply but that's the case with any

system because if if ethereum's market cap goes up by five billion you got five billion dollars of USD going to like crocken or other exchanges and it's the same thing was not quite the same thing you know because the ethereum price goes up like then you know everyone's richer and like you didn't you didn't have if the ethereum price goes up and now it's 5 billion higher okay there's look some marginal buyer that will like bought at that higher price but you don't have to you know expend all of the money to kind of account for the increase in the market cap yeah that's true that's true but like I think the thing about Mel is that um anyway like if you actually read the paper while I describe it it's not

exactly that you have to Mint out every single M it's more like this minting process establishes a price feed for how much does a market value M and if we need to print more mels the protocol actually prints more Ms uh you don't need like the amount of minting doesn't increase substantial in that situation like the minting people only indicate to the protocol that hey Mel currently is overpriced print more mels or Mel is currently underpriced Stop Printing so that's really the only thing that it indicates and then the protocol prints more M and it sells the M yes and it sells a male in exchange for

the other token or sa token yeah right so basically in the case of now the M Supply increasing a lot uh does it mean let's say now it goes from Z to 5 billion then some of it will be I guess spent by people doing the sequential computation and getting the mail but some of it will be people basically buying mail or like buying the Sim token and exchanging it for mail exactly yeah so then it sort of also means

that uh a lot of the value of the growth of the Sable coin or you know the the male market cap would be sort of absorbed by the Sim token yeah exactly so the Sim token is overall designed as the value of cruel token so it also is where all the fees go to um when you stake Sim and red validators you get fees that are all denominated in Mel as people use M so the more people use Mel like the more value your occurs to so tell us a bit more about Sim what are the sort of properties of the Sim token oh that's a much more less interesting token it's just a standard fixed Supply um proof of stake token um like there's

nothing cool going on there and the reason why I separated it from Mel is really just because what's good for a proofer stake token is different from what's good from a curreny because for proof State token you wanted to be like a crew value you want it to be fixed Supply you wanted to be like if I own a third of the network I always own a third of the network um with money you don't want them so it's kind of like you know um shares in the bank versus dollars right and that's kind of the idea so I mean a lot of proof State tokens are inflationary right where basically paying block rewards to the to the validators right for you know running the infrastructure so if it's a fixed Supply token you don't do that

here or like yeah don't do that here because if you think about it right like inflation's not free money inflation's you're really opposing attacks on everybody else who's holding the token and um you can just Express that as a fee you don't need to like inflate the token um so of course you need to design the fee Eon fee economy differently but Mel's whole design ensures that like this the work that the you know validators put in can be fairly compensated purely by transaction fees and I think that's a more transparent design it's also like more clear where the money's coming from and you know it's also like just like more free

market right because with inflation that's a variable that you know you need to centrally tweak it might be too high it might be too low the market can't correct that because it's baked into the protocol with fees it's different right like if the demand goes up the fees goes up if demand goes down the fee goes down and it all scales as uh with the actual social value of the Mel protocol you just have to design the right mechanism so that the just captured not and I think I have successfully done so it's quite a different way we charge fees from other blockchains um but yeah cool um anything else you want to talk about I don't know there's there's a lot of other things right like for example how like a Brendle really ties into

Mel's overall go to market and how that'll work right because I guess I touched onto that right like Gaff will turn into an a Ral frontend but I think like the bigger thing there is that once you have a lot of users on a rendle then you have a ton of demand for anonymous traffic relay on the air rendal Network that means that you and the listener if you run a node you can actually get a lot of those um a lot of that demand as revenue and this is also really cool because what this means is that a rendle gives you a free market for satisfying people's demand for the free internet and I also Envision a world where this

means that internet shutdowns and internet citizenship becomes impossible because if if let's say you're some country and you shut down the like International internet then the price for transferring air rental traffic in and out of the country would go up like crazy and if you have like a starlink or like a little cable across the across the border you're going to make a ton of money by serving the air rendal Network so we're really what a really does is that it uses the Invisible Hand of the market to solve internet censorship and I think that's the thing I find the coolest about a rental cool well thank you so much Eric for coming on and and thanks for walking

us through it I think that was really great and I think gave a really great overview I do think this is one of the most yeah just like Regional projects right so of really trying to think through things from uh from a sort of you know fundamental design perspective of like what we actually want to do with this and then yeah there like so many original things you came up with here so it's very exciting to see where this is going to go yeah definitely cool if people want to get involved what's the best way to do so well join our Discord go to our website Mel project.org you can find all the links to the right things there so if you join our Discord uh we have a lot of live Lively discussions about reseearch

and our protocol development just join the conversation cool well thanks so much Eric it's really great to have you on you too