In Conversation with Vitalik Buterin, Justin Drake and Karl Floersch (Ethereum Foundation)

Scaling: Ethereum’s make or break challenge.

We talked about this backstage but What’s the animal on your t-shirts this Is a buffer corn a buffalo unicorn with A nice little F logo on the top okay I Have the same shirt as you and even I do Not know that well yours is even more Crazier because it’s even more animals Right interesting so let’s talk about Scaling I know a big jump of subjects Yeah I guess so So I looked at the agenda of the meetup Ta ta Meetup Mhm and I liked not one not two but Three different keynotes And panels about skating in one way or Another it seems like it’s like the most Important thing on your mind right now How to scale atiim let me start with you Italic from an outsider’s point of view It feels like a TM got really popular Because it was so versatile you could do So many things with the protocol but is It the same reason and the same thing That is causing these skating issues Today that is that it is so versatile Again ultimately right like price is a Function of supply and demand and so When transaction fee is on any of these Block chains whether Bitcoin or Aetherium or whatever else become high Like that can like that basically does Mean that there are a lot of people who Are who are sending transactions and who Want those transactions to get onto the

Blockchain ins and well the blockchain Is popular enough and people are Interested enough and using it that League they basically have to compete to Get those transactions actually income Included into the slots and the blocks Right so and if you had to design Material from day one today in 2018 Would you have made it a bit simpler to Make it a bit more scalable from day one So there’s definitely changes that we Know about now that we could have Definitely made into the protocol from Day one that would have made it Significantly more scalable though And there are also ways in which the Problem just is fundamentally hard and The things that we’ve learned over the Last two to three years of seeing the Ethereum blockchain wife have Contributed a lot to our understanding And to like our ideas about how to make Blog chains better in the future so Let’s talk about these changes one by One because I it’s pretty public the Road map and what you’ve been working on So Justin maybe you can start by talking About charting a bit and proof of stake And give a bit of an overview and Explain how the casper fork would work Right so initially the two big Initiatives of the foundation from a Research perspective were casper the Proof of stake and shouting and these

Were kind of distinct and they were Going to be launched separately and also Be separate designs the idea of proof of Stake is partly to provide a new Superpower to blockchains which is Called finality the other idea of proof Of stake is to reduce the cost of Consensus and for shouting well that’s Pretty easy it’s just scaling up and I Guess through the research that has been Done the last few months we’ve decided That actually it might be best to Combine the two in a single design Partly because there’s lots of common Infrastructure and they would able to Make a more seamless design to integrate The two and more cleaner design a Simpler design but it turns out that There’s also things that sort of network Effects between the when you combine the Two so for example with sharding you Have all these different shots and like Parallel universes pretty disconnected From each other and so you want to wait For the shots to be able to communicate With each other and we do that by what We call cross links or checkpoints and In order for these cross links to have a Strong strong value and that you can Lean upon them you want these Checkpoints to be finalized And so this is where you get value from Casper and integrating the two means That you’re both security and

Scalability basically right yeah and so Will it require multiple updates or will You combine everything in one single Update Casper charting so I guess we’re Now calling the oval endeavour of Casper And sharding FM 2.0 and this is a very Ambitious project which even though it’s A single design it will likely be rolled Over several over several years my Expectation is that Casper will come First possibly next year 2019 and then In in sharding we’ve actually broken it Down into two big phases phase one is Actually the the data layer so coming to Consensus as to what data is in the Shards and then phase two plus is all About the states is about giving meaning To that data having a notion of Transaction and my expectation is that Phase one can come you know after Casper Maybe 2020 and then phase two 2021 maybe So it still years away but you’re not Just working on that we also have other Projects cow maybe we can talk about Radin and plasma so many cryptic Cryptocurrencies and block chains are Playing around with the idea of creating Side chains payment networks and payment Channels to offload some of the work Outside of the main chain so raiden and Plasma a bit different but maybe you can Give us a bit of an overview and I’m Interested how would the two work Together

Sure so we just talked about Casper and Sharding and how that will kind of give More security and more scalability to The main chain but those technologies Are actually built using like Fundamental tools like crypto economic Mechanisms That we’ve been working on for multiple Years and the there are many purposes For these crypto economic mechanisms and One of those purposes is actually Scaling up gaps as they exist You know if as scaling up depths that Can exist in today’s etherion blockchain Right so you can actually use things Like Merkel trees to you know provide Commitments to a large amount of Transaction data so that is actually Plasma so first we’re talking about Plasma you basically have the ability to Commit to a large number of transactions And not actually send those transactions On to the main chain instead you just Provide this crypto economic commitment And then later on you can prove that you Own a particular coin by proving that You have you know received the last Transaction in a large chain of Transactions so it allows you to like Scale up and then things like raid in And state channels you can even have General-purpose state channels are these Unanimous consent agreements where Everyone’s signing messages and

Essentially by acknowledging each Other’s messages they’re able to update State without putting it on the main Chain so essentially the idea is if you Use these mechanisms in an intelligent Way it’ll allow you to actually provide Guarantees about transaction inclusion And provide the the services that we Want to provide with blockchains without Having to wait for the underlying Scaling solutions however the underlying Scaling solutions do allow this to be Even easier because you can actually now Send you know for instance plasma Requires regular block submission to the Main chain and when we have a main chain That is much more scalable that means That that block submission is going to Be much cheaper so there are kind of Synergies that go along but the Important thing is that these Technologies are not holding you back as A decentralized web 3.0 engineer this is Something that you can do today and That’s super exciting is it fair to call Plasmatic crypto kitty’s updates you can Definitely do crypto kiddies in a way More scalable way this is like the one Of the key insights right you can make Crypto kitties by deploying a smart Contract on the blockchain and every Time you meant a new crypto kitty you Just send a new transaction and it Deploys a new smart contract or

Something but you can also do crypto Kitties by you know deploying a just a Single transaction hash which contains a Merkel tree of a whole bunch of crypto Kitties and then you can breed them by You know submitting Merkel trees and you Can do interesting you know use Cryptographic and economic tools to then Actually provide scalable crypto kitties And that is like super exciting and and Italic I know you’ve been are you Excited as well he’s a good seller I Have to give it to him Vitalik you’ve Been tweeting about gas prices and and Because we’re talking about cryptic it Is what was you reaction economic Reaction when you saw that the network Started to be so congested because of Crypto kitties and because of the you Know massive uptick in I mean I think to Me I just highlighted that no demand for Using public blockchains Yes hi and we like this screen needs to Really step up a step up our game in Terms of increasing the scalability in Order to be able to meet that demand and So okay you need to up your game so Let’s talk about governance for a minute That’s the Italian community have the Right structure to pass all these Updates so so one very important point Here is that sharding and Kasper are Updates but plasma is not an update State channels and right and are not

Updates right now they’re not Forks Plasma and state channels are layer 2 Technologies that can be developed on Top of the existing ethereum blockchain As it exists yes and the people have Already been slowly starting to test and Implement a state channel and plasma Project there’s no lightning Network on Bitcoin there was that liquidity noticed Thing that came out a few days ago There’s right in funfair did you like Fake channels which is their fancy Marketing term for state channels Yesterday saying that many people are Using plasma as a buzzword for something That is not really blessed now yeah like There’s a lot of people that seem to Have the Impression that plasma just means take Your centralized server and shove a Merkel tree on top of it and that’s like Not what it is magic unicorn technology Yes basically yeah but it feels like Bitcoin is a bit stuck in a corner Because it’s like a cryptocurrency with No head and and they don’t have a way to You know pass some decisions together as A community Bitcoin has plenty of heads you know you Have the core development team then you Have the miners with associated pools Having 53% of the hash power can they Agree no no I mean you have gridlock and Exactly you have gridlock and and do you

Think at some point you could face the Same kind of issues and the same kind of Gridlock that very very possibly Definitely and it’s so this is part of The reason right why I think the Etherium community very early on has Been kind of like very you know both US And other etherium developers and people In the community in general have been Kind of continuing to kind of promote The idea of the etherium roadmap and Look if basically if I feel like in the Eyes of most people in the etherium Community the roadmap is kind of part of The social contract and that includes The idea that proof of stake will be Done that charting will be done that There will be much more efficiency that There will continue to be upgrades going For some time into the future and people Who buy into aetherium of basically Buy-in knowing that these are the thing The things that you know people are Going to what the people are going to Want to push forward and so when people Join the community and start building Applications on it they can know that You know they are doing this on a Blockchain and inside of a community That is kind of like dedicated to a Continuing progress so there may be dead Walk on you know what what specific path The community should take but I do feel Like we’ve made the very important step

Of Basically committing not to stagnate and I think for a platform that’s still a Rapidly evolving technology in a rapidly Evolving industry that’s a fairly Important commitments to have and ok Let’s talk in a very pragmatic way let’s Say we are in 2020 sharding Kaspar maybe There is some plasma technologies that That are available around how will it Affect gas price and how do you think Miners are going to react so plasma Clearly competes with the etherium being A base chain for transaction inclusion Right so basically what that means is That application developers have a Choice where they can either build their Application in a way that uses a lot of Base chain transactions or they can Build a plasma chain into a lot of Plasma chain transactions so that Clearly reduces demand for the base Chain though it increases the demand for Transactions that are willing to pay Very large amounts to get onto the base Chain because they’re very important Like the plasma Commandments and plasma Challenges I mean I expect in general Don’t make gas prices go down very Significantly is if we see a fairly Widespread adoption and though I expect It all ends up like expanding to a Little usage even more ok interesting so You talked about plasma and the finality

Of transaction and how security is Important maybe just in you mentioned Security as well do you think with the Current roadmap you have the right Approach when it comes to security to Make sure that everything is normal and Business as usual so from the the Sharding perspective the security kind Of comes from pooling all the validators In in the same place and then randomly Sampling the validators and putting them And the shots from the point of view of Of Casper it’s actually an extremely Simple protocol and we can just write Math proofs about it there are just a Few lines long and they give you strong Guarantees about things so for example If you have two inconsistent finalized Blocks then you have a guarantee that at Least one-third of the validators will Get slashed so that’s what we mean by Finality it’s actually economic finality In the sense that if something goes Wrong then a lot of money will be burnt It can recover by itself basically right So that’s kind of separate right so that There’s also mechanisms for recovering And mechanisms for making sure that you Don’t do a bad things don’t happen in The first place so Cal when you’re Working on these new updates and and Technologies like plasma how do you test Your code this is definitely a difficult Kind of process because you a lot of

This stuff is like very kind of Fundamental research and a lot of it is You know mathematics that we’re trying To we’re trying to basically analyze This these protocols this these attack Vectors in many different ways and I Think the key thing to kind of ground Everything is this adversarial thinking So we want to create mechanisms that are Both efficient but also that adversaries People who want to kind of take Advantage of other users or do anything Wrong that doesn’t really do what you Want it to do those people get punished So we just talked about how one third of All the validators get slashed if They’re to do something that we Basically don’t like and so that is one Example of how even if there is some Kind of you know failure mode that Happens there are consequences for those Failure modes and so that’s what we’re Kind of what what this process of Building these protocols looks like is Okay we we think up you know what do we Want to achieve what guarantees do we Want to provide how can we penalize People if those guarantees are not Vided how can we reward people who are Providing those guarantees and then once We have this kind of objective we come Up with you know as much math as Possible to describe the scenario but of Course these are complex systems and so

We also need to like simulate them and We and the simulation tools are really Weak right now there’s a huge space for You know people creating you know agent Models that kind of like simulate Different actors economic actors Decisions and come up with kind of Really complex game trees that that we Can like use to further analyze if we There’s like a a missing piece in our Actual protocol but it’s basically this Like iterative process of you know come Up with some good ideas you know try to Formalize them simulate it implement it Realize that it we’re wrong go back to You know our ideas we you know and Repeat and repeat and repeat interesting Vitalik do you work with hackers to try To find security holes before pushing an Update I mean uh yeah I don’t like the Word hacker because there’s all of these Like science-fiction connotations they Don’t actually quite my life yeah well We have C we have security audit Researchers on our team we have a bounty Program we get disclosures all the time You know it’s totally all part of the Process cool and let’s talk about the Foundation for for a second so you have Structured this whole community around The the etherium project how do you find These people to join your team basically Choral hell have you been found well I I Found this stuff on the internet

Watching Def Con zero YouTube videos and So vitalik and Vlad came out with all of These really really cool you know crypto Economics videos and I was just like Hooked on them and so I read up Everything I could and just kind of Contacted vitalik at a conference and Then was like oh hey can I do work for You and he was like sure and then I just Started doing work for him and You know he was like okay do this now And I was like okay I’ll do that now and He just kind of snowballed into me Working on Casper and sharding and Plasma and so but now thankfully that Was the kind of the the sparse days now Now there’s the III search where people Can kind of look at the research that We’re doing everything we do is public And put out there in front of the Community there’s also crypto economics Taught study where there’s a learning Tutorials and stuff like that for for Learning this kind of crypto economic Stuff that I’m working on um and there’s Hopefully even more videos and Presentations and Justin Drake has an Awesome edie con presentation for Instance and many more to come so you Have a lot of energy I have to say so After having heard that how many people In this room wants to talk to vitalik And say hey I want to work with you That’s my main question maybe you

Shouldn’t have said that on stage Because now you can get flooded with Applications and people coming to your Door do you guys all agree when you have To take decision how does it work Internally sometimes we agree sometimes We don’t agree and we both think for a While and eventually we find something We both like so usually there’s a third Way the first way or the second way yeah Vitalik is pretty amazing and his Ability to actually like compromise if You believe something justin and i had This moment where we didn’t really like Vitalik sharding design and we kind of Iterated by ourselves for a while built Up our case any yeah and it’s been very Public do you think it’s healthy it were These discussions online in front of Everyone oh yeah how do you look at Other block chains do you think there Are some implementations that could be Interesting in that you could work with To put into the etherion projects There’s definitely some that have that Are doing a lot of interesting Pioneering work in cryptography that we Follow closely and you know we’ve I Think we’ve been very close with the Zero get with Z cash community for Example yeah And that’s the only one um there’s Others saying we’ve talked to what said You talked with Definity people a couple

Of times we’ve talked with some there’s A there’s a couple of other blockchains We also talked to kind of academic Research groups that are not I never Filled with any specific blockchain yeah And then we’ve had close contacts with Ic3 for a couple of years like the Stark Where researchers like people at Stanford so when you when you go all got Into this field did you think you would End up being being some kind of Researchers because maybe you engineers By trades and but now it feels like you Researchers well I was initially a Bitcoin magazine writer by trade no I Thought I’d like okay this is a cool Idea Just like crank out some Python code Implemented for a couple of months and Then I go back to university and like Somehow that ended up not happening Today involved with I don’t really do a Production code as much I tend to focus On like prototyping because that seems To be where I provide the most marginal Value Bullock yeah it feels like it’s a Different work to work on prediction Code because you need to put code that Is so efficient and easily available Mean the other thing is production code Needs to be written on like fast Languages and I like Python so you all About Python basically maybe you should Have put a Python on your t-shirts to

Defend interesting all right let’s talk About other updates in the pipeline so You mentioned recently idiom to dotto And and and it feels like a big mouthful But are you already thinking about What’s next after that what would be Next so we are thinking of FM 3.0 as Well okay Part of it is reliant on fancy Cryptography so one Things we want to do eventually is have Our system be fully quantum secure and Minimize the assumptions that we make Photographically and one of the best Tools for that is called stocks which Are like stocks but there’s no setup There and they also quantum secure now We’re also looking into things like Erasure codes to make data availability Proofs and potentially things like super Quadratic shouting but that’s still very Much research topics and do you think Are you hopeful that you’re going to Push these updates against quantum Computing before cand quantum computing Is here yeah I think yes very like Oh With a significant margin like what like What to be secure computing I think is Actually less hard than a lot of people Think like basically because you just Hash algorithms are all quantum secure So you just kind of make everything Based off of hash functions and you kind Surprisingly enough in a lot of cases

What do we can do that and I have one Last question because unfortunately We’re running out of time and did you Think about opening a t-shirt store to Sell you to show designs I know other People did like Vitalik where’s calm or Whatever like whatever autumn figure it Out okay so you let it open for Karl Open-source designs basically Interesting all right thank you very Much [Applause]

You May Also Like

About the Author: admin

Leave a Reply

Your email address will not be published. Required fields are marked *