The Invisible Shield: Protecting Blockchain Users From Themselves (and Everything Else).

Transcript

[Music] Hey guys. So yeah, nice to see so many people. Uh yeah, so the entire purpose of my talk is basically to wake you up. So by the time Vitalik talks, you are in the best mental shape able to digest everything. This is just a joke. you don't have to wake up. So yeah. Um, so I'll start asking you guys a question just again to wake you up a little.

Raise your hand if you know what product market fit means. Okay, some people know. So it's actually it's a very simple thing. It means that if you have a startup then uh you have to iterate your startup until you get users and actually preferably paid users and if you don't do this your startup basically dies and the actual term product market fit comes from the internet bubble times. So there was this internet bubble in 1999. People had lots and lots of money, even more money than Ethereum ever had. And then there was a crash. The bubble crashed in 2001. And

suddenly people didn't have any money. And startups actually learned things like four out of five startups die. You have to do something. You have to find users. At that time, Google almost died. So it was pretty tough time in 2001. But then uh things appeared like uh lean startup there was y combinator people started to learn that actually if if you get users and preferably paid users you don't have to have so much money. So the entire different philosophy was born where if you have a startup you stay frugal you save money and you basically change things until you get users.

So if we look at u uh bitcoin for instance and ask the question does bitcoin have a product market fit the question the answer is yes and the reason for this is that you know the value of bitcoin is to to actually store the value of money and to save for retirement. Bitcoin is slow which doesn't really matter for it because it's supposed to hold lots of lots of money and gold is heavy. it doesn't like hurt gold in any way. In in the same way, the fact that Bitcoin is slow doesn't hurt Bitcoin. So, Bitcoin does have a product market fit. But, uh if we let's take a look at web three,

right? Uh another question to you guys, raise your hand if any of you have a web 3 application on your phone now and you are using it. Okay, I I don't actually, but you guys have. So there there are some early adopters but you don't really see people in in public transportation in subway playing web3 games you don't see people using like defy much so all of this stuff is uh still very experimental I think and it's a little sad because this number here says eight right so we're like at least eight years spent and we don't really get so much adoption and so

the question is why Why is it happening? Why we don't really have so much of this web three thing? And there are three things which actually are required which are missing. I would say number one thing is privacy. If you pretty much take any web two application and you're trying to port this application to web 3, it's impossible to do. Even in a simple example, if you have a bank or if you have an e-commerce store, there are private things, right? If you order from Amazon, then your order is private. The money you pay is private. The same with your bank, right? You don't want to show people uh what's inside of your bank account. And Ethereum doesn't really

have privacy. And because pretty much all other blockchains since Ethereum just copied Ethereum basically, uh there's no blockchain which has privacy. And we've been a little bit stuck for the last, I would say, eight years. Uh there's no privacy, there's no security. Security means that people feel safe. Like the hack uh that happened like just several months ago with this pure poor guys that like lost a billion dollars, send it to somewhere else through uh gnoses.

This would actually never happen with web two. You know, you don't hear people in a regular bank being sending like a billion dollars and not getting this billion dollars back. So, the security is missing. And the reason the security is missing is because in regular banks in web two, security is provided by humans, right? Ultimately, if something is wrong happening with your bank, your credit card gets gets, you know, blocked and and then you call the guy and the girl and they actually unblock your credit card. So the security is actually achieved ultimately by humans and blockchain doesn't have humans right. So on blockchain people feel unsafe and then last thing is agentic. This is kind

of a interesting word but what it basically means that a blockchain is passive you know it doesn't really do anything for you. If if you have your bank your bank can send you can have a Python script and the Python script can send you send you an SMS message. So to authenticate all kinds of stuff if if you have a trading account you can run Python script on your trading account to trade all of the things are missing from blockchain again because blockchain is passive and because pretty much everything copied as Ethereum doesn't have it so no one has it. Uh blockchain is decentralized. Uh so we basically have this thing where web two have has a

product market fit but it's centralized and web three does not have a product market fit. People just don't use it. You know it's still experimental. So most of the web two applications you can just take a very simple one just even like a web store. It's impossible to port it to web three. So in in effect web three doesn't exist much and at scale we've been working for the last five years as layer 2 for Ethereum uh and uh we have lots of transactions. We have about three million transactions per day. We have almost 20 layer layer two chains which are zerium compatible chains. But at some point we started to feel that we essentially got stuck. you

know uh there's a constant set of problems this privacy security and agency things every time we talk to a customer or someone would want to do web three they had this problems and people were getting stuck so we kind of number of years we were trying to kind of stay the way but ultimately there was was lots of discussions in our team and we decided to kind of try to fix this problems essentially go our own way, our independent way. We still love Ethereum but uh we just want something more exciting I think. So we decided to start our own layer one which is called fair. We actually have

it running now. The the phase one we started the internal test net and gave access to the test net to our partners several days ago. So we're running this this layer one and the entire team is kind of very excited about this. I'm going to talk about fair and what we're going to do with it. Essentially it's ethereum plus and the three things which we want to fix is privacy, security and agency. And we just want to fix them in a sane and and fast way. We don't want things to go forever or like have a plan for the next 20 years. We we want things to be solved reasonably fast. And so we need to solve the three things and the

privacy we are solving using encryption something which we call proof of encryption. Essentially fair is the first blockchain that is able to hold encrypted data. All other blockchains just have public data. We have encryption. It's it's pretty hard to do but I'm going to talk about this a little later. We have we're going to have security. And security means that you somehow need to match the humans that exist for for web two. there are no humans. So pretty much the only way you can do security for uh web 3 is AI and essentially even ch2 could probably prevent like 99% of uh existing attacks like the guys that send

a billion dollars to the wrong address. This would be easily pro prevented by by ch2. It's not so difficult to avoid human errors for the AI. So we need AI and we need agency and I'm and agency there are several ways we are going to do it but uh the realistic way and fast way to do it through trusted execution environment or something which is called SJX. Okay. So first I'm going to talk about privacy. Uh uh we have four bas four phases for for the protocol something which we call bite and uh I take the credit for in inventing this uh backron name bite blockchain integrated threshold

encryption. I did it using chip. So, and chip came up with this pretty nice acronym, I think. And it's kind of a more mathematical acronym. We're also calling it proof of encryption. The idea is that you want to actually encrypt things on blockchain. But then the question is how do you encrypt? Because on blockchain blockchain is not a person or it's not a single entity. So there's not like a public key of a blockchain. Turns out there is an encryption algorithm which is called threshold encryption where you you essentially encrypt to a group of people or nodes uh something actually very similar to

multi-IG but multi-ig is for signatures and and threshold encryption is for decryption. Basically what you have is that you have this uh jigsaw puzzle. So you encrypt a public key but when you decrypt you you need to have a super majority of nodes in the network two/3s of the nodes and each of them decrypts its piece of the jigsaw puzzle and then two/3s of them get together they they glue together the pieces and then they actually have the actual decryption. It's pretty cool process and means that you need to have if you have twothirds of the network honest any single or small group of bad validators malicious validators cannot actually you know do

bad things to your encryption and decryption. So we're using this for privacy and phase one of of bite protocol is actually probably eliminates uh me which means that uh when you submit a transaction uh the transaction is encrypted. It doesn't actually break the ethereum compatibility. It's just a small JavaScript library. you encrypt the data field of the transaction and then basically the people don't see the transaction until the transaction is uh in the blockchain and just before it's executed by the EVM the transaction gets decrypted using threshold decryption it's pretty cool process and actually

you probably you can prove that there's no MV so you take any smart contract you take unis swap you take a you take things like polygon like prediction markets and you don't change any smart contracts, the blockchain stays the same, but things magically become uh magically become uh MV free because the transaction is in the pending queue in the memory pool encrypted and by the time it's getting decrypted, it's actually already in the blockchain. It's impossible to front run or analyze or MV whatever.

Uh yeah, so that's how it works. It it's a very simple slide actually. Basically, you use Ethereum as your kind of you used to use it. Uh but then there's this as I mentioned this like small JavaScript library and you encrypt your transaction and then it goes through all of the Ethereum stuff and then we have a provable consensus and the provable consensus actually at the as the last stage just before VM execution decrypts the transaction. And it actually sounds easy but you need to have a very detailed specification for the consensus to do it because you know that the problem with consensus is always always it can get stuck. It also can get kind of decynchronized. You know the state

can be different on different nodes. So you taking a consensus and you plug in this another pretty complex thing which is threshold decryption as the last large stage last stage and it doesn't have to break it but you have to have for this really really mathematically defined consensus and we were lucky that we're using consensus which we kind of compiled from academia and all of our stages of consensus are mathematically defined. So we were able to to actually plug in the threshold decryption part and then it's magically magically MV free. So we are working now with uh our partners and friends to essentially port MV defy applications to to this new chain to to fair and if any of you guys

are interested we'll be happy to work with you for on this MV free kind of MV free protocol. Again, this is actually like uh I was trying to use CHP to create an illustration for a protocol and then we have a really talented engineer and he said, you know, this CHP is really ugly. I'm going to draw it and he drew this. So, it's it's pretty cool thing. Uh it's probably hard to understand every single detail but the basics is that you have a zerium transaction, you have the two address, you have the data field and you encrypt the things and then they go through all of the zerium mechanics through the consensus encrypted and they actually the transaction is committed to

the blockchain encrypted and then the last moment the EVM decrypts the transaction. It actually works. So there's this life cycle as I mentioned you submit you encrypt to painting you encrypt it and then ultimately proposal is finalized committed to the chain and then at this last moment this last stage uh you have execution by the EVM before execution you actually decrypt you use the scripts called decryption a committee committee decrypts using over the network using a super super majority and you have lots of interesting use cases for ME free blockchains.

Immediately you definitely have MV things like unis swap. Uh by the way actually there was a really great talk before this my talk about me. I can only say that MV is one of the reasons we have so bad adoption of decentralized exchanges. Most people use this centralized exchanges because just the exchange rates are so so much better and they are better because on centralized exchanges you don't have me extraction and centralized you do have it. So and also you have lots of spam on chains coming from this me guys. So but in addition to exchanges you have lots of other things applications of any resistance like games uh prediction

protocols whatever. Uh now this is phase one is already running. The phase two is something which we are working on now and that that's way actually can decrypt you have encrypted data you can decrypt on demand. So the smart contract can say say I here's the data and the smart contract say says decrypt and this is how we will do it on on on fair in solidity. Basically there will be a pre-ompiled contract and you will just pass the data at any time and it will get decrypted and then why why would you want to use it? You know, this is kind of a computer science exercise for some of you guys that are interested. Why would a smart contract want to decrypt things, right?

And the the easiest thing is actually the rock paper scissors game and it's just for you to think about this. Why would a rock paper scissors game need to decrypt at the same time? The reason is that when you do this like rock paper scissors, you actually reveal at the same time. So the smart contracts essentially connect collect stuff and then and then decrypts at the same time. It's super important and cool. But the real application is more for from for things like defy because with encrypted things you can have encrypted bids for instance for for auctions. So you can have auctions where people submit encrypted sealed uh bids and then at the same time atomically they get decrypted.

Now lots of other applications bite protocol phase three and that's where we are actually adding the reenryption something to share data that's really important for things like social networks for instance you have your data on chain but uh you want to share it with someone else without sharing it sharing it with the like entire world. So this is this cool thing which is called reenryption. You don't decrypt it, but you add the public key of this new person to to the encryption keys and then this new person can also decrypt and you can do things like Google Docs.

This is how it's going to look in solidity. And then you can do things like Google Docs. You can do literally like web to like social networks where you share information with people you like like Instagram, maybe Twitter. I think on Twitter everything is public but Instagram, Facebook. uh and then cal collaborative document editing like uh Google Docs for instance when you when you share documents uh and lots of interesting other things you have this confidential gaming pretty much with public blockchains we have now it's impossible to do any any like reasonable game if you like play poker you don't show your cards to other people right so with private data you're able to fix pretty much everything we

need we need private data you fix things like you have an econ hammer thing and then you want to encrypt the orders. And the last phases which is the hardest is the uh doing mathematical calculations for things like onchain banking and this is the hardest thing. You need to use things like threshold fully homorphic encryption. And there you can do things like you have a bank, your data is encrypted and then you you want to do like a tax on on this data taxes and then you can selectively do a mathematical computation without decryting the data.

Lots of interesting things here. Now this is the encryption proof of encryption thing. The second thing about product market fit is security. And this one we are actually solving by adding uh AI uh agents to the blockchain. And there are two important things. Smart contract deployment. When you deploy the smart contract, the decentralized AI can actually analyze the smart contract and assign the security score. And then transaction execution like the guys with 1 billion thing. When you execute transaction, AI can actually look at it, look at the score of the smart contract and talk to the user to actually explain to the user whether transaction is secure or not secure. Actually providing this like a trusted agent. Last thing is

agency. Yeah. And agency we are actually doing using uh trusted execution environment and AJX. That's it pretty much guys. So lots of use cases we are working on on fair. You know all of you are welcome to join our community. Uh we are moving we going to launch this layer one also move the existing scale chain software. Thank you so much. [Applause] [Music]