Agreement is Its Own Incentive - TOKEN2049 Singapore 2022

so I'm here to convince you that agreement is its own incentive now we're all here because on some level we're super excited about blockchain right and why is blockchain so exciting because it gives us three things that seemed kind of impossible 15 years ago and the first and the one that's gotten the most press of course is coin distribution with blockchain you can create a new virtual coin out of thin air distribute it in such a way that there's a limited Supply but people believe it has value right this is a shocking uh maybe not anymore but certainly 15 years ago who could have believed that Bitcoin could shoot up to you know sixty thousand dollars at some point second of all blockchain gives you the ability to have censorship resistant transaction submission so with apologies to our previous speaker if you build your business on top of something like Facebook or apple or Amazon uh and you're too profitable you run the risk that that company is going to come and say hey nice business you got there how about we take a larger share of your profits or we're going to shut you down right and if you build on a blockchain you know that nobody can do that to you because it's it's decentralized but finally the most exciting thing about blockchain is the fact that it provides secure irreversible transactions even across unknown and mutually distrustful parties so this means that you can build like basically anyone with a text editor can build and deploy products uh on a blockchain that would previously have required really hard to get bespoke agreements with kind of trusted intermediaries like Banks and whatever so this is how to just an enormous multiplier effect on the pace of fintech innovation and we've just seen this unbelievable amount of innovation now if you're going to have a blockchain you need it to be replicated on a whole bunch of machines run by a whole bunch of different people uh around the world right and that requires a consensus protocol for the machines all to agree on exactly what transactions have executed in you know in what order and of course consensus Protocols are a long-studied problem in distributed systems but there's a new twist when it comes to blockchain because most blockchains use incentive-based consensus right so we're all familiar with the concepts of proof of work and and proof of stake uh what these really boil down to is the fact that we can grant virtual coins to people as a reward for doing things that make the consensus more secure unless making everybody's transactions harder to reverse and the reason this is uh this is so interesting is that it lets absolutely anybody wants kind of anonymously participate in administering the blockchain and that open membership of the administration helps to guarantee the censorship resistance right because you know even if there's one Miner who doesn't want to publish your transaction you know eventually somebody else will include that transaction and finally of course consensus the incentives for consensus are also kind of uh an ideal mechanism for coin distribution right and that's how you know Bitcoin became the the Juggernaut that it is for example but what I'm going to argue today is that you know incentive-based consensus great for cryptocurrency for many kinds of tokens you're actually better served by something that we call proof of agreement than by either proof of work or proof of stake and specifically I'll argue that proof of work and proof of stake are really not what you want uh if you are uh creating what I call issuer-backed token so this is any kind of token where the value of the token derives from some kind of promise that someone has made off chain right so this could be a cash back stable coin it could be loyalty points it could be tokenized Securities it could be carbon credits things where the the token itself is tied to something in the real world and the reason uh incentive-based consensus is not so great for this uh is that you kind of have this Goldilocks problem if your token transactions are too valuable for the underlying incentives that are securing the blockchain then you'll have insufficient security right even ethereum which is super high quality blockchain has maybe like 20 billion dollars of each stake right so that means you could kind of mess with it starting at 10 billion dollars worth and if you could lever up maybe you could you could borrow 10 billion eth for a lot less collateral and so on you know a few billion dollars you could really mess with ethereum right so does that mean you shouldn't use ethereum no but if you're a central bank and you're thinking of of building something that's going to grow to a you know a trillion dollars of stablecoin maybe you don't want to issue that on ethereum because the incentives won't be sufficient right there may be nation states that are willing to spend a few billion dollars to disrupt your economy another problem with this incentive-based consensus is that it fails to leverage a lot of investment that we've made in society in mechanisms for reaching human level agreement so these are things like the mechanisms we have for agreeing on ownership rights for enforcing legal contracts for regulating Banks and ensuring they stay solvent and can't walk walk away with your money and finally maybe the kind of scariest thing about real world tokens that are issued on a proof of worker proof-of-stake blockchain is if you have a if you issue a billion dollars of stable coins say and the blockchain forks you may now face two billion dollars worth of redemption requests and obviously you can duplicate the blockchain virtual assets you can't duplicate the underlying dollars so what this means is that you know on some level all blockchains actually require human agreement right you have to agree on what consensus protocol the blockchain is using what the parameters are for that protocol what transactions are valid what the semantics of those transactions are uh in order for the underlying consensus mechanism to be at all useful so as an example in 2016 ethereum changed their transaction semantics to incorporate this weird one-time irregular State change that was designed to recover funds that were stolen by uh exploiting a bug in this contract called the Dow right and this could have been a disaster because not everybody agreed this was a good idea about 15 of miners said you know hey too bad about the theft but the rules of the rules you know we should continue mining under the original rules um but fortunately the ethereum community handled this extremely well and even though the sort of minority who didn't like the state change uh continued to mine on the original blockchain they called that ethereum classic and everyone kind of agreed that ethereum had the new set of rules right and less catastrophically more recently uh the merge switched ethereum from proof of work to proof of stake and everybody kind of agrees that ethereum is now proof of of stake and the old system is called ethereum pow or efw but the point is that if we didn't have human agreement on on what the blockchain is what ethereum is then the underlying consensus would be useful for it would be useless for security right imagine coinbase and binance don't agree that the merge happened and one of them still thinks that ethereum is actually ethereum pow then you could send a token like usdc from One exchange to the other and that token might get lost because it gets sent on one blockchain and not on the other right and that would destroy credibility the blockchain industry and it would actually significantly hurt both companies so there's a really strong incentive uh for everybody to agree on the rules and what ethereum is right uh even though usually that human agreement is not actually formalized right another example you could be you know embracing the future and put your apartment lease on a blockchain but if someone breaks into your apartment you're still going to call the building manager or the police and you're going to want everyone to agree that you're the rightful occupant of that apartment and so that means agreeing on kind of what the what the blockchain is so uh the point is not that this is bad to have human level agreement the point is that sometimes you kind of need this human level agreement in order for the the blockchain to be at all useful and so given that you need sort of a higher level human level agreement why not use that as the basis for your consensus protocol you can do that without kind of expanding your attack surface and this is what we call proof of agreement which is the approach taken by the Stellar blockchain right and there's one big caveat here which is that this doesn't work if you want coin distribution right so we're not trying to replace Bitcoin here what we're trying to do is provide a better alternative uh if you want to issue tokens that are anchored to some kind of real world asset right and the way this works uh is that basically the validators in our system are not Anonymous they're run by known organizations with a reputation and uh every validator every server that's running the software in the system um is actually declares these are the organizations that I have to agree with and I won't reach consensus unless these other people agree with me so I have a little example on this slide where we have three validators at the top V2 V3 and V4 and they all agree say we're not going to reach agreement unless all of us agree to the same thing right and you only reach consensus when uh everybody's been mutually satisfied so for example validator V1 here says well I won't agree to anything unless V2 and V3 do and then uh then uh that means that by the V2 and V3 are saying we all won't agree to anything unless V4 does so by the time V1 reaches consensus you're kind of guaranteed that V4 is going to agree to the same thing obviously the details of this are well outside the scope of this talk but we published this paper in the 2019 Symposium on operating systems principles so uh if you're interested in the technical details I'd refer to that paper you can get it at stellar.org sosp 19. so the advantages of this approach are it's ultra low cost because there's no work to do for proof of work and there's no staking required either and so that means that we have uh really cheap transactions with very low latency right um and uh another Advantage is that token issuers can actually protect themselves from double Redemption tax if you want to issue a trillion dollars worth of some stable coin on stellar that actually makes sense to do that the only thing is you have to run your own validator or contract with someone you really trust to run the validator and then you tell people this this stable coin is worth a dollar because I'll redeem it for a dollar but when I process those Redemption requests I'm using this exact validator V1 so if you care about the value of my token make sure that you also agree with that token and the mechanism we use for this is called Quorum slices so basically you put my validator in your Quorum slices and you'll be guaranteed that you'll never disagree with me so Stellar uh with this consensus protocol this proof of agreement consensus protocol has been in production use since 2015. uh I've done over 5 billion operations in that time to give you an idea uh last quarter we saw about 38 million payments on the system uh and about a quarter of a billion dollars in relevant asset transaction volume and by relevant asset we mean assets that we believe are incredibly tied to some underlying real world thing like like money and uh again because there's there's no mining or staking or anything the transaction fees are incredibly cheap right cost less than 100 000 of a dollar to do most transactions on this network and the consensus protocol runs in well under a second on average so uh it just kicks off every five seconds to avoid uh growing the blockchain too quickly uh but that means that the latency is kind of equivalent to what you expect for a credit card swipe uh and of course our energy consumption is extremely low because there's we're just not doing much work here Stellar is also the blockchain used by MoneyGram and so in advance of this is if you have stable coin on Stellar you can actually turn it into physical cash you can cash it out at MoneyGram locations in like 200 countries right to get your digital asset back into a physical asset so there's an annual conference of all the people building on the Stellar Network it's called Meridian taking place in a couple of weeks in in Rome uh and so if you're interested in attending or you just want to stream some of the talks you can check it out at uh meridian.stellar.

org and if you want more information on the blockchain you can get it from the Stellar development foundation's website so that's www.stellar.org thank you