Running EVM on sidechains: five innovations

Transcript

okay my name is Stan Klitschko I'm the CTO of scale Libs scale network so first of all I'm gonna just tell you guys a little bit about myself I'm pretty much like a typical nerd so I did my PhD in physics in Germany at Max Planck Institute I moved to Los Alamos National Lab in New Mexico did physics there moved to Stanford University a Department of Physics also did some physics research at some point I moved to startups as I get as many people moved from academia to startups and I took a course from Professor Binet who was a crypto professor at Stanford really liked the course and at the end the top actually joining a startup that Professor bond has started and this

startup ended up raising like 50 million dollars and we were building a crypto processor hardware crypto processor we ended up selling the start of exactly for like 50 million dollars so we didn't make any money but it was really a great start of crypto career and then I found that the crypto lab which was affiliated with National Institute of Standards and Technology in Maryland in the state we were doing all kinds of crypto evaluations for for the US government so I looked at this lab for several years then I found that a Wi-Fi security start up for this startup actually got back to Ukraine we had an engineering team in Ukraine and then ended up selling the startup to a company in London and then after that

startup we got funding for any AI / machine learning startup which was supposed to simplify AI and machine learning and we're the top building the deposit the prototype and like piloting the product but was really hard to compete with Google and AWS of this world so the starter was like dying pretty much and so they were like cut back to investors and said you know I will like just kill this this startup my big rehire and but at the same time it was 2017 so one of our investors from Rick Thompson from Signia ventures is like what you guys have to go this thing is zero and they had no clue what it was you know

I knew about Bitcoin but I didn't really like it so much but when I found out about the serum I was like well it's interesting you can program the same so I literally jumped on the airplane and flew to DEFCON 3 and the moment I like entered if country in Cancun it was like flower power like Summer of Love everyone was like dancing and metallic was dancing on day on day on the stage this unicorn and stuff and I realized that even the deceased don't give me money for this I have to drink the thing because there's no Google and there's no Microsoft it's really like wild field so get back to San Francisco and then I actually got together with my co-founder

Jack he had the successful healthcare software startup and he was also working on digital currencies so he raised money from VCS and the initial idea was to do like take our big data thing and speed up a zero and the first idea was it's an F s decentralized exchange but then we thought ok maybe instead of doing the exchange we do the infrastructure that textually we'll be able to support this exchange and so then that's how we got to this idea of scale pretty much a parallel network which speeds up a little but then has to be involved actually we realized that transactions weren't really the thing for 2020 that people don't really want to pay for

coffee now using Bitcoin as zero maybe they will in 2023 but first startup is really important to find a you know the timing so then it kind of keep on thinking and realized that the thing is a decentralized cloud and we in the top building a community so we're trying to take this small fish and put together all of the small fish that to create this community and it kind of a big fish and this big fish you want to compete with AWS essentially essentially creating a decentralized cloud so essentially what you want to do is to take every server on AWS and turn it into a blockchain and it's very simple and that this blockchain is going to be th compatible you'll be able to

use all of your it is tools for this essentially anything you can run now on AWS you'll be able to run on this on this network on schedule but for this one we do need to kind of tweak EVM a little bit tweaked solidity a little bit so we're making some small extensions it's still compatible what we're extending it a little bit so people can take a website move to zero you can take mobile app moment to zero woman we're really practical we I really want you guys this liked in the year of 2022 to be able to do the things and then let me just give you a really fast poll so what do you think the mankind needs more decentralized Facebook who think the mankind needs the decentralized Facebook

please raise your hand okay who needs who think the mankind needs a decentralized Oubre raise your hand Wow many people taking uber is like how who thinks a mankind needs the decentralized AWS and to ask the question who thinks the mankind is the dista flies google Wow Google is like yeah and I can tell you guys my own story so so for Google it's interesting several years ago I was in in Silicon Valley and just some random ways I got to an entrepreneurship party and there was the co-founder of Google search Bream and people who just like approaching him like drinking with him

and asking questions and I got to him asked this question I said do you believe that at any point in time and open-source search search engine can actually compete with Google and he said no an open-source Association doesn't have the resources we have millions of computers from pompon you know open-source doesn't have it guess what maybe actually a depth of time thinking the cryptocurrency provides to us this ability for open-source to have actually a financial some kind of a financial you know business model so actually can have a decentralized Google and I actually if you don't could think about Google you know I can tell you another story running this Wi-Fi start-up and our

startup was doing this freemium model so we're doing this public Wi-Fi access points totally free like you could have up to 28 access points put on our cloud and manage them for free but there was no solution and and is anything like that but the Google Google was still putting us on the third page and to get on the first page of Google you had to pay five click five dollars per click to get on the page so your tiny startup having great solution free solution there was nothing like that it's except like Cisco which I was like super expensive Cisco was on the first page we want the third and to get to the first page you had to pay five dollars per click so when the god has a start up into this situation I realized that how much Google Stifler's innovation how

much this a entire idea of Google how much conflict it has because they say they're the best but the best search engine in fact actually they can't be the best because they want to charge money to put you on the first page if they put you on the first page without charging money they will lose money so so what you see you have this centralized entities all of these companies the startups Google Facebook you work when they start they're very innovative but at some moment they develop a brand and they start like extorting like they're like corporate and power and that this the problem we have in the economy now we have too much people like going for the power and to less people too few people going for

innovation so we have lawyers and great guys but maybe we'll have like too many lawyers in the syrup maybe too many cups too many soldiers and too few innovators entrepreneurs and scientists and a blockchain provides us a way potentially to change the thing and maybe not not the perfect thing I'm not saying that we are gonna like change the world and revolutionize I'm much more like an evolutionary guy and just want things to become a little bit better I'll be happy about that so that's what we do at scale we create trying to create a network where you'll take be able to take Facebook this year move it to zero take AWS moratorium and they're very practical and you're trying to find like the simplest solutions in many cases also find compromises

that's nothing massive project so so actually that's what asked on Twitter on Twitter also people are like talked about Google tip like don't like Google I said so I'm building a decentralized cloud we want to pretty much you'll be able to take every W server transformative blockchain kind of now this is a big thing we don't say that we're gonna do it over like one month but it's like if you miss Tesla you know when Tesla was started there was Tesla Roadster and then Tesla Model S and then Tesla Model 3 so they did read this the great marketing that they were actually able to keep people for many many years with this idea and that trying do the

same we're not saying that this entire transition for decentralized cloud is gonna happen over over it like one year it's probably gonna take 20 years because what's gonna happen is gonna be different generation and as with Google you know in 1998 it was clear that Google would become this huge company but it took them 20 years to become this monster because they had to raise a generation of people that actually researched Google the first thing very much with decentralization you know I think it's gonna take lots of time but we need to start with something so the network we are releasing this year it's pretty much our Tesla Roadster is the minimum thing and it's probably gonna for the most part target defi because

defi is the most kind of sexy thing now people actually doing loans to each other over over Internet and it's very exciting because the previous idea if you think about huddling it's not really the mainstream you know some people are fine by the by Bitcoin wait 20 years become a zillion year but that's what what most people want in this life most people want to have money put money somewhere just earn a little bit more interest than they make in a bank so this defy but the great thing about it it crosses the chasm in a sense that now yes crypto is available just to people you know so we're thinking about defy like major thing this year and then what we also see social networks I kind of picking up people talking about since

like social networks you centralized uber you know all drivers really don't like uber you were take 30 percent of their money but if you're a driver you have to pay for gas you have to pay for repairs to your car for this 30 percent if you if you compare to how much money you take home you can actually let's like double your salary if there will be no 30 percent so we are thinking organizations like that can actually really make it come any more effective and also make goobers and Google's maybe also more effective because they'll at least have some competition okay so that that's our idea it's kind of long-term vision and we want to create this met this community of small fish running this decentralized cloud on this decentralized cloud people

will run this work chains all we teach compatibility interacting with each other and every blue chain would be essentially what the server is now like a Facebook or Google you pour whatever so we I'm just talking really fast gonna go like like about five things we had to tweak the VM because we have to be able to run a website on the VM or a mobile app right so one thing virtualization we take like our guys the guys the Quran our network the validators they buy things we don't want text you them to buy an AWS virtual machines we much better want them to buy servers the Carver servers so buddy they pay for the silver now we split the server in small pieces

Suton docker containers maybe some of you guys know docker containers and the idea is that you'll be able to buy a very cheap blockchain because this year the minimum containers had one over hundred twenty eight so you'll be able to pay very little amount of money and have your own decentralized blockchain that is executed by people for you the community so this our architecture pretty much split the server many pieces every piece can be part of a separate block chain so we call it elastic block chains now file storage this is actually interesting story you know when Steve Jobs was already very ill and pretty much dying the guys who founded Dropbox they he wanted to actually acquire Apple wanted

to acquire Dropbox and the guys came to Steve Jobs there was a meeting and seems up like offered this guy's drew and there's a guy who found that Dropbox his before offer them like X billion dollars and the Dropbox guys like no we're not that take this X billion dollars want y billion dollars right and Steve Jobs said you know I'm not gonna pay you this wide billion dollars because it's not the product file storage is not a product it's a feature you know it's not something that needs to exist by itself and that's actually what we also think this for this ipfs but our take is this file storage has to be part of the blockchain it's a feature or not and not a product so extended EVM you can store any files on it you can store images

music anything you want and we've got like many hackathons now and people putting all kinds of stuff on a zero competition watching so they easily host the entire like website or using like images and music whatever so that's the storage part not another thing is consensus so we took pretty significant time to understand which consensus to use and we started like trying to categorize and understand which ones are secure which not secure so we develop like a framework like so I do it basically we can we can take a consensus and split it in like seven stages and every consensus is split in seven stages very much the same way if you have a car you can say okay this car has break this car has you know wheels

you know engine and things and if you are able to split your car into smaller logical components your budget you're able to analyze the car better so now we can take a consensus from someone else and say okay this guy's great it works very fast but that's not have brakes you can't really break so it's not realistic your car so we have this thing and so we have this high level thing which is mathematically clean and the rest of it all of the pieces all the seven pieces we have we actually steal from academia we didn't design any of it we just looked through different academic research and found different algorithms that people in academia we were actually perfecting and we are implementing them and actually putting them together especially for consensus we have a great algorithm which has been implemented at University of Nance

actually here and friends and I think that's the best algorithm you know in my subjective moon what kind of it's 2015 and has lots of very good scaling properties it's mathematic algorithms but anyway so we have this consensus because it's a new thing it has very modern things like there's no block time the block can be really fast if you you're under the corporate net fear of the block time could be less than 100 seconds and there's no fixed block time and also there's no time outs in the consensus you can have like a set of machines you can unplug all network cables and then wait for a week and you plug in the network cables again and because there's no timeout and anywhere in the code so our consent was for class on synchronous Norfolk's book time Provost a consensus mathematically

provable from the beginning to the end and one of the things they're gonna have this year is actually published my paper and submitted to every free journals actually there is a dependence review of this algorithm people say okay this secure or maybe a security hole somewhere whatever so we are still working on that one and because it's mathematically provable we know that there's a solid foundation if you have a bug in questions we know it's not like the foundation is bad it's like somebody like put an incorrect code and actually makes a lot life much easier because if everything is mathematically provable your bugs tend to become smaller with time and less important while if you look at the foundation you kind of trying to fix things on the fly you can

have a huge bug every day or huge crash you don't really know how how to fix this thing last number two paranoid compatibility visit eh maintenance so you can honor stuff you can pretty much use all of the ETH tools you can using like truffle whatever open Zeppelin anything you you can use on the main methods it's fully compatible and so these are the tools we prank guys probably know and then we also provide a very simple way to transfer tokens actually a new messages token suggested type of message transfer any tokens like a receipt one-year 777 and ft you can like transfer back and forth for these chains like mini chains and then and then the token transfers everything is

an instability so we kind of we tweaked the EVM but for developers if you have some to smart contracts on two different networks the way they interact is just by sending messages to each other in solidity don't do anything you need to know is solidity for this so that's how the messages are sent we have and then lost by at least is decentralization this is kind of the the hardest problem really currently for us because we start the network but the network the community is a complex community in a bit like Bitcoin is simple people like use peww they mine and then they submit blocks in whatever our stuff is necessarily complex we can't really make it less complex because many things like

EPS you know storage anything so so because we have complexity when we started the network we have to be ready for this thing to get embarrassed and crashed and like brute blue screen of this our hope is that the community will help us and create the circuit the our validators will help us to actually perfect this and this is gonna be I think a huge huge undertaking for us because we really want this small fish to actually grow into this large fish and all methods are fine I don't think it's going to be just cryptic I'm not economy because many people are just slashing because for instance for slashing not everyone likes it you know people get discouraged when they the money is the technique taken from them in the community kind of gets trained so

so we'll I think for the most part it's important to also create some kind of a psychological or social wasted for this community to exist and that don't really believe in like this community by being anonymous I believe in the things more being more like cooperatives like relic guys get together and create XE company and the taxi company work so guys like this guys get together and create a decentralized cloud so this are gonna be a decentralized cooperatives of humans and one huge thing which is totally unknown for us now how to organize this humans so they actually provide enough service to compete with AWS because it doesn't work well we can't really compete this AWS if we push them too hard they will start leave leaving if

people get upset but you know they get penalized too much the history we will have to find out how we create this this this community and it's a super complex interesting problem you'll see how it goes don't really know how it's gonna go we hope it's gonna go well and I think the best is going to be just iterating and iterating and and share and kind of changing things so we have delegator czar talking stella gettable so some of the people may not just delegate their token that they won't know how to run the things but they will delegate a token to someone who runs the thing very much like if you have a small taxi company to run every car I have to have like whatever $100,000 and someone needs to delegate to me this hundred thousand dollars to place a proof that I'm a good

guy to run this taxi so we have delegated and validators and they have different different things in the network but as a delegator you will be able to buy very small amount of of tokens and delegate to one of the validators now we actually face this huge problem about scalability because some of the validators may literally get like hundreds of thousands of delegator delegations delegated to them that's really interesting thing by the way we have many many interesting things on our website and blogs and if you guys want to read more broad network we have lots of interesting information okay there's a time line so I guys told you pretty much we have the Testament now we are

working on releasing be main it and then the things with the first iteration of our network will still have not have or every single thing we need but two things which will still be missing one thing is AI if you have a social network you need to be able to somehow analyze photos when people upload them otherwise you know if you want to have a decentralized Facebook you should be able to somehow filter bad things on the way in the community should be able to and Facebook does it by using AI so you'll be able to actually upload in our second version you'll be able to pro load any tensorflow model to zero and actually the sensor flow model will be able to analyze the photo but then you know guys tensorflow models can do

anything they could have lots of financial things one of the really exciting financial things you could do with tensor flow and neural networks is to do a stable coin where the fluctuations of the coin are smoothed out by a neural network hopefully more advanced than just a regular mathematical formula so that's really the lots of interesting financial applications of AI on blockchain and private computation we do need to be able to give people reasonably private data and we are implementing this using Intel SGX essentially in the second iteration of our network you will be able to have your private things that you only only you will know and they will be protected by Intel for many many commercial applications is going to be fine

maybe if you if you like store a key to some government and say whatever thing maybe as rakes it's not good enough but if you store your banking your home where you paid for coffee or which restaurant you did go to if you think it's reasonable so that's it I would like to think our investors and if you guys have any questions if we have time I'll be able to answer questions yeah we're kind of running like out of schedule I guess okay great thank you better sorry oh wait is John here yes