Hashing It Out
Hashing It Out

Episode 105 · 1 year ago

Hashing It Out EP 105- Optimism

ABOUT THIS EPISODE

Today Corey talks with John and Ben from the Optimism.io project, a scaling solution called an "optimistic roll-up" for Ethereum. We're going to go into how this project works and what we can expect in the near future as well as further down the road.  

**Links**:
- https://optimism.io
- https://twitter.com/optimismPBC
- https://hashingitout.stream 

Welcome to hashing it out, a podcast where we talked to the tech innovators behind blocked in infrastructure and decentralized networks. We dive into the weeds to get at why and how people build this technology the problems they face along the way. Come listen and learn from the best in the business so you can join their ranks. All right, welcome back to the show, episode one hundred and five, I think. Yeah, today we're to talk about optimism. You're on hashing it out. I'm Dr Corey Petty, and got a cohost guest, John Mardlin, who not only does the show with me but also works for optimism, and we got been Jones, who works for optimism, so we run the whole gamut. I guess people know who you are, John, but why don't you go and introduce yourself and then we'll all ben and then we can talk about what optimism is. What's up, everybody that's looking for you to cue that up? Oh, yeah, it's okay. Yeah, I have run and John Marlin. I also go by merillion in many plays online. I am the sometime host, Co host with cory of this show and and I work at optimism and so I'm that's grey said. I'm sort of sit in between. Maybe I'll serve as a bridge and something Ben's not saying. Our bends and saying doesn't seem like it's like I can help them it down for the clubs, the audience, if I think that's necessary. Hey, what about you? Cool, what's up? Guys? I am one of the many folks out in the world named Ben Jones, but I guess the reason I'm on is because I've been working on this optimism thing in some form and other for about about two years now. So, yeah, it's a pleasure to have me on. Thanks Coy, thanks John, for sure. Let's get let's get the out of the way. What is optimism? Oh, great question. So optimism is a public benefit court that is the current steward of the optimistic etherium protocol and that is a scaling solution for atherium that we're building. It basically puts etherium inside of itself in a way that is, you know, a roll up. So it maintains a lot of the security properties of the layer one but makes it bigger and better and faster. So you said a few words there that I don't think most people here at least in like the conversation of scaling solutions and Roll Up. So I most often. But this is this conversation got started and it moved away from plasma. It was like roll up sort the new scaling solution for Etherium, and that's going to be a variety of two flavors. Either there zero knowledge roll ups, which chooses which is dealing knowledge proofs to then post stuff on chain, and the optimistic roll out, which does in a different way. Optimism by its name, it's a kind of tend to the idea that it is an optimistic roll up. But said a few other things. Can you elaborate them at a bit? Sure? So, yeah, so, actually, my my background was working on a nonprofit scaling research or called Plasma Group. That's a lot of the folks sort of carried over into optimism and I guess the thing that I would have beat by saying is really there's scaling solutions that are optimistic and Zek and the delineation within that is actually, I think, a little less clear cut than everyone thinks. One of the early posts out there that I would call out is one by vitality called the dawn of hybrid layer two's which was basically one of the first posts that was that sort of started to really gain attention in the roll up space, and so it's worth calling out that these plasma things and these roll up things, they're all sort of related under this one framing of you don't execute the transaction on chain unless there's a dispute about it, and this is sort of the formula by which, in any of these scaling solutions of the optimistic variety, and by the way, there's optimistic plasma and zk plasma, this is how you this is how you get to those things. So, yeah, I guess I'm forgetting the initial words that I said out of is radomly on, because he said plasma and I wanted to talk back to that. Is there anything particular I can dive in? Hope you out on. Yeah, so you're calling at a court now, which is probably a reasonable thing to say, because the only time you actually executed on chain is when it's disputed and then there's like, I guess, consequences of forcing the layer one chain to execute it. And so like I kind of want to dive into the deep of what differentiates optimism from the other scaling solutions and one of those kind of pros and cons are. HMM, okay, Great. So first off, like why did you switch naming to like a court, like people was a pcb or PBC or where it oh, public benefit, Corp, corporate...

...out. So, yeah, so, okay, yeah, I can start there. So the reason that we switched over was because we knew that sort of this stuff was exiting the main research phase where that needed to be a focus, and it was time to implement, and so we wanted to do that in a way that wasn't just a traditional startup that was gonna, you know, just have you know some of the incentive problems that we might see there. So we landed our public benefit court as somewhere in the middle. Basically means that we have a project charter. That is, you know, US being able to keep things open source and promote our mission, which is sort of democratizing access to to blockchains. You're saying Corp, as in Cor I'm saying Corp. Oh, yeah, I'm just talking about the the corporate structure there. Yeah, okay, great, okay, great, that's it. I don't start about that a bit then, because that will green to dive into a lot of technicals. So like what is what? What is the corporation and why is it necessary in order for because I like the I say public benefit corporation is fuck automatically gives me feelings in two different directions. One is this is like this is something that's a public good and so it should be shared by anyone who wants to use it in a corporation, which is like I'm going to take all your money as best I can. So I how do those two think make? Of course that's not that's somewhat of hyperbole but like, how does this work? How does this legal entity actually insert itself into the development of this technology? Yeah, Public Benefit Corp isn't actually more on as a hilarious joke that I hadn't thought it before, but it's really not one. So I mean, if you go online, you can go to optimism that I own. Read our charter, and so basically every public benefit corp has a charter that it has to uphold, and so ours is something along lines of, you know, promoting access to open source products and, particular, particularly blockchains. What does it mean for us in practice? Well, I think the main thing that it means, like, along with the general vibes of open source is just that the goal for optimistic etherium is to become a globally maintained software and so it's very important for us to not have incentives in place to retain control. So starting out you need a bootstrap, you need to pay people, you need to get the core software written, and so that required some components of traditional startup. But once you're there you need to be able to relinquish that control, and so basically we have that there in place to be able to do that. You want maybe talk about then briefly, like the idea of exit to community and and I think like the importance to the team of avoiding things that might lock us in, as you know, centralized controllers and the system is something that like is a calm, constant conversation that I think is really interesting. Yeah, so that's you're totally right, John. Like one of our you know, our most important goal is to in when of o we track our sprints and when we're going through our engineering cycles of pushing out, you know, releases and developing code. Our main goal with that is we have a community section where we go over what is the pull requests of the community. Have contributed and like our ultimate goal is to have the number of points, the like amount of work that went it's from the community to exceed what we can do, and this is like a core tenet of what of how we've designed our software protocol to basically be a fork of existing etherium projects that have massive amounts of contributors and, like John Said, everyone on the team is very hyped about that too. So yeah, basically we know that the only way for a crypto protocol to succeed is do what atherium has done, which is have everyone working on it from all sorts of parts of the world under every all sorts of conditions, and the that pace of build on development is just going to outmatch anything else that can happen. So we're really like focused on making sure that one we have a codebase that's going to be accessible and familiar to that group of people to be able to bootstrap that. That's why we use all these theorem technology and to it's why we develop everything out in the open and and push for things on that front. Kind of think about where I want to take this conversation, because a few like core concepts that I think I'd like to try and get a grasp on, especially for the audience, because I think there's a few bubbles of misunderstanding, or not misunderstanding, but like curiosity about how things are going to go moving forward, as I ad as like me, as a potential user, would actually use this stuff or develop this stuff or supports like supporting infrastructure for this type of stuff. Right. So, like I run a bunch of notes. Personally, if I want to use the optimism network, this this layer two network, that there's no thiss, basically this court system of like passing around, passing on, executing transactions and then optimistically playing them, throwing them into the layer one.

Right, do I need to run additional notes? What do I what do I need to do to just support it? Like how does this thing actually have infrastructure that it's going to be completely outside of a theory, know, because its own network that then that interacts the etherium? WHO's running this network and what type of resources as it cost to do so? MMM, so that's that's a great question. So, first of all, yes, it is a distinct network and of course the importance nuance there's that it's bridged to the main etherium chain, and so that's the the part of it being an optimistic roll up is sort of how that bridging structure is define and the strength of the security of that is what, you know, makes it different from, say, aside chain, where you also have to run notes. But in terms of the functionality that users can expect, it is very similar to its own network in that you do have nodes that are running an optimist you know, that are optimistic etherium nodes, as opposed to l one nodes or, you know, nodes for some side chain or like polygon. And Yeah, so that's how it works. So it is a different thing and obviously the mechanism by what you get in and out is like depositing and withdrawing, so you bridge between it. Obviously, you know, we have support for infrastructure providers, of people like Infura and stuff like that. So you have a similar experience there, where power users can run their notes if they like. Other users may not choose to and the system progresses. Of course. Additionally, if you're running the node on Ltwo, you're playing this system of watching l one and compared to what's being done there and potentially prepping fronty disputes that might surface. Obvious that the obvious question and feeling from from people. This is coming from from me too. I've partly know the answer, but it's something that you're going to get a lot of. Is How does that differentiate itself from F to, because that's exactly what I'd be doing as a note operator for F to. Hmm Oh, so it's very similar. I mean, I think this is why you see a lot of the like phase one and done or vitalis roll up centric. If you're in a future is that we're realizing that a lot of the things that especially et phase to those sort of execution layer for sharding that is planned. We realize that a lot of that can be done with roll ups and it's an extremely compelling way to go about things. So first of all you do is complimentary because it gives you more data to post to the roll up, which is still ultimately one of the potential bottlenecks for roll ups, but you can actually do the execution without having to wait for sharded execution in layer two is and it's basically naturally shartered in and that way. So yes, it is very similar and just like e too, you should be very excited to run a note a running as me does as it possibly can. But, like, I think this is the there's a somewhat of a breakdown in relaying this type of stuff to anyone who's not a developer or just been an ecosystem for a long time, and that's basically with like the naming of what these things are and how they differentiate themselves. So, like, if I think about F too, that's somewhat of a transitioning network from a theorium one it, but ostensibly it does a lot of the same things it I run. I run a separate piece of software. It watches the etherium one chain, whether that's an atherium or Theorem one note that I read myself, or I'm relying on some other centralized infrastructure like like Infura and based on and it's it's it's chained in a way. It's bridged to the theory in one network, and that the validators, the valuator pool comes from the the deposit contract on a theorem one, and so it makes it. It has to watch that to make its own decisions. Now where the differentiation happens, at least with respect to etherium too, and it's current phases and all of the layer two stuff is what it's watching for and the data it pushes back and forth. Can you talk a little bit about like what an optimism node is actually watching and what is it pushing back to layer one and getting security back right? HMM, great question. Great Question. So there's two note roles to differentiate their actually and the answer is going to be a little bit different for each of them. So the first thing I'll talk about is a verification. I guess the first thing I'll talk about is what are both of those notes watching, because they do watch some things in common. So basically they're watching for blocks that get a rolled up. Hence it is a roll up. So what is I'm and in practice? Basically, to be able to dispute transactions, you need to guarantee two things. One, you need to guarantee that any transaction in question which could be disputed is accessible, it's available and that that's important, because someone could propose the outcome of the chain is x and that could be wrong,...

...but if no one had the transaction to disprove it, it would be impossible to demonstrate in a dispute what the actual transaction is. So the first thing you need to know. You need to be watching for is transactions that get rolled up, and basically what this means is you submit a very large transaction which does nothing other than like Hash and notarize all the transactions which might come into question down the line and stores just as very cheap hash on chain. So this is where you hear the layer one, is the data availability layer, come from. Right, it's because we need the reason that we need to make those transactions available is because disputes are useless if you can't always find the transaction that you were disputed. So that's the first thing that you're watching. As a note, and, by the way, those comes in come in two forms really. They come in the form of deposits or sort of l one to L to transactions. You might think of the mass, which is in effect almost always a deposit, as well as user transactions that are just with transactions within the role of an art communicating between chains. Okay, so that's one thing that you watch. The other thing that you watch is basically a set of proposals. So you download those transactions, you look at the bodies of those transactions and you run them your no, just runs them right, adds them to the Etherian chain and you then get a state route. That is, they effectively what the result of all those transactions are. Right, be that you arc twenty balances, be that you know, some ask whatever. Right, it's a smart contracts execution and the other committed to that. That's a state route and those are also being submitted to a theory. So you watch for those two things. You Watch for the transactions and you watch from the state roots and if you see that a transact, you by watching the transactions, you know what the state roots should be and you basically check there for dispeak, for potential disputes. What I do? That? Oh, great question. Well, I mean one, because you want to say that you're the first person to ever submit a challenge on a name that roll up. That would be insanely cool. Beyond that, of course, like it is a in a Benette positive action in terms of submitting the challenge, like you can get rewarded for doing that. At least you recoup your gas cost. Fundamentally, like, why do people run notes? Is a is a different answer for everyone. So Corey might be advancer for you, then it's going to be for you know, other major note providers. Lots of people are getting paid to run notes and that's going to be an end, you know, a market that exists on layer too, I think, possibly even more so than layer one in the long run. And there are also certain applications and power users of the role that want to run their own notes. And at the point you're running, you're running the note. It's free to be checking for dispute, so you're going to do that. But one example of this is, well, there's a whole kind of worms there, but you can do liquidity provision where you basically provide faster withdrawals as a sort of and this is one of the prime these cases of people that definitely want to run their own notes. So that's that. was that no one of the two types of notes correct? That's right. That's right. So those those you can think of as like we call them verifiers. You can think of as them as equivalent to like the full notes. Then we have a secondary, second type of note, which is currently run by optimism, and that is called the sequencer, and the sequencer is basically there to collect user transactions, to aggregate them into blocks, order them and submit them to as the things that are being watched by the full notes, and so that's kind of more equivalent to the minor role in layer one or the validator role in these too. And so that sequencer basically users can send their transactions to it and the the sequencer, will bundle them up and submit them on chain. So two things I want to ask about a sequencer. One, I guessaid if I've found I like to do a mental picture. I see the secreencer basically doing most of the processing of a lot of the stuff and the validate and knows just kind of watching, taking in what they can and proced and doing stuff as they need to based on whatever their business case is. But ostensibly the the valuator nodes or the kind of watching nodes are there to make sure there's like one people aren't cheating and to the sequencers and sheeting exactly right. So that's that gives me like a good mental picture of like a swarm of watcher nodes doing whatever business case they care about on on optimism and the secretser node funneling a good portion of all these transactions so that they actually make it into the main chain appropriately. Is that a good mental model? Yeah, I think that's a good mental model. Yeah, absolutely. The other thing that I'll call out those really cool about seat about sequencers is you can get basically your state confirmed faster than it's laid down on l one, which is a really cool property. So the sequencer actually maintains, like, at in some simes, an extra optimistic state of what it is going...

...to submit. Think of it as like the next pending block right. It's like I fart to think about, like I'm try to like. I mean if you think about it in terms of almost like a hierarchical tree of of evm like block chains, you have one layer of security in the sequencer, which is like the MEM pool, is all the transactors flying around, it gets confirmed to the sequencer, and then you have the time it takes for the sequencer to submit blocks in the layer one chain, which is like an additional layer security. Exactly, exactly, okay, and then you have, I would you said previously, which you kind of neat, is that Peepe like liquidity as a service is people saying, will screw all that, if you trust me, and then I'll just I'll just give you whatever fund you want on either one of those chains and I'll handle that weighting process for I'll wait. I'll deal with the waiting time for a small fee. Yes, even better, you can do with smart contracts, so you don't have to trust those people. Doesn't matter. Even better, let's trust even better. So I'm very interested to start mapping out risk is this thing starts to solidify a little bit. But, like one of the other questions I had for the sequencer is, I'll get to the I'll get to their security part later. What are the resources of running a sequencer and how do those scale, particularly with the number of people who run them? M that's a great question. So one thing that I want to have beat is bought. Sort of definitionally, you only have one sequencer at a time, okay, because the sequencers the person choosing the sequence, and so you two to that's like a leader likes out whatever. However, that grows in terms of number of participants. That's a leader based protocols exactly. Thanks. Okay, yeah, then then the answer that scales obvious. It's whatever. Whatever the actual when work is for any for submitting a block. How does that go from optimism running the e what a how do you see that scaling computationally? The commet how do you how is that going to be able to how the how's a hardware going to scale as you need to handle the scale of the number of people submitting transactions? That's a great question. So definitely we have a big focus, an optimism on vertical scaling. So there's this this trivial way, obviously that you can get more through, but where you just deploy another chain, right, and that's like horizontal scaling. Well, we are of the opinion that the single chains that we see today are not sufficient, and that's not because the we're going to just increase the hardware or cuirements forever. We have to have big breakthroughs there. So that's a big focus of our work, I think. As it relates to your question, I think I think what I will I think what I will say is that what we are now really seeing and what, at least from our perspective, layer two is really started to highlight, is that there is different computational requirements between validation and block production, and I think this is something that's like a little bit overlooked right now because the gas, for example, for those two constraints are tightly coupled on l one today. But one of the things that we're a big fan. Up is increasing the hardware requirements for that sequencer. This proportionately to the hardware requirements of the the verifier, and you can think of that as there's a great tweet by Jeff Coleman where he calls this first run gas pricing and second run gas pricing, where the where you actually have different properties, that it comes that it requires to calculate a block of transactions and likes determine whether that's valid versus playing it, playing it later. Anyway. I'll pause there. But but the answers yes, and the coolest thing is to be able to increase those hardwarequirements on the sequencer, make them be big beefy, you know, service providers, without doing the same thing to verifiers. It's kind so. So this is interesting. I haven't heard this idea before. It can makes it, thinks. It sounds like a cryptographic cash function in a certain way, right, like the compt like, where valid verification is much, much easier than like the initial computation. Yep, and the main source of this is going to be parallelism. So basically, if you have you know, you can consider an extremely simple version of parallelism where you just take a block and you out of validity condition that says a block is invalid if to transactions touch the same piece of etherium state. Right, so they both touch the theium or the UN swap. You know eth price. That's an invalid block. As soon as you make that a block limited condition, then at then you can take any valid block and you can paralyze it and you can run it on multiple course at once. The trickiness is that for the sequencer that is not so easy, right, because it's hot, because the nature of the Etherian virtual machine is such that a transaction could, this is the money Legos, right. A transaction could execute the unit swap trade at any point in time and you and it might be the last thing that it does...

...at the end of a very long piece of computation. And so that first run of figuring out whether or not transactions can be paralyzed, it's going to be very expensive and so much harder problem. But once you've done that and it's just a blidity condition, then it's trivial for the verifiers to go and split that up. And I have so many ways in which I can take this conversation. Okay, let's look we'll take it once, one step at a time. The next one is the like the obvious one, which is going to be a lot of those minds. How do we how does it go from falling? Does it go from being a optimism run verifier or like sequencer to a community run set of nodes that take two turns being the sequencer? If so, what does that look like? Great Question. So our earliest thinking about this was about a year and a half ago. We put out a post called MeV auctions, which sort of said, hey, we have this separation that people aren't making right now, which is between inclusion of transactions and ordering of transactions, and that Med is clearly showing that the ordering of transactions is very popular and this should be called a sequencer, so on and so forth. Since then, thing, I haven't thought about it that way because I was having a I was having an argument with someone in the Bitcoin podcast lack, around the concept of Mev and what it would look like to have fairness built into the sequencing of transactions while maintaining like openness and permission this at the base layer. So, like, I'm sorry, just I want to introduced that that I had that and I didn't think about it from this perspective that way. So like, okay, so are you saying it like the sequencer is a form of any of the auctions? Like this is, like this network was basically a way of sequencing transactions that eventually end up on the on the main chain, whereas or you can just go to the or you just go to the main chain because, like I said, a Mev or like the ordering of the transactions is a market in itself and it's proven to be incredibly fruitful for those who are able to take advantage of it. Now right, and it's very interesting. It's a very interesting question. What is the market for ordering transactions? So just a rewant for a second. The the thing that I was originally talking about. There was sort of our first stab at understanding this and a first sort of, you know, construction, and the idea was sell off the ability to be the sequencer, basically just have a recurring auction or something of that nature. And what is important to highlight is that that is not the same as what we're seeing with like flash pots going on right now. I don't know how much context listeners have sold up. Be careful with that, but really the market for transaction ordering today is mining. Right, it is still the miners who are determining the ordering of transactions that they've become a direct link to. How I get my introduction and where I want to is going straight to the miners and bribing them to put it where I wanted to be. Right exactly, and and well, and really you're not going. And really, when people are paying for that transaction ordering, the estates right, they're usually prying like flash pots. And really what flashbots represents is a reseller of transaction ordering. So it's the miners that get to decide the order, that ordering of transactions, but it's then flashbots that's going and resawing that to them, because miners have a sept there's a separation of concerns where miners are really good at mining, not necessarily ordering, so they can outsource that to a market. Just to heads up off off site. I'm hearing every time your fear hand hits the desk that's coming. It is the not talk while doing it. I do the same thing, but I can hear it. But yeah, so, like, okay, so that's that. The it's being resold because it is a tremendous amount of work to optimize the value from ordering transactions, and miners are spending all of their effort just mining, because that's that's its own separation of concerns. Is staying up to date on what the Best Hash is for place improove work. That what the Best Hashes for given block? Yes, okay, clearly. So how does that move forward for you guys? So what it means for us is that we have to we basically have to observe this market condition and figure out how that is going to inform what happens for up to this you through. So basically, what it tells us is that if there's somebody that has the ability to sequence these transactions, then it's in there. It's probably in their economic interest to be reselling this because even in the case where you are a even in the case, it seems that the people that are determining these orders of transactions are no longer just one party, right, it's a group of parties. So Flash Bots is like has this bundle aggregation where there's different parties that say, Hey, this ordering is good for me in this way, because I have this specialty, I know how to deal with this liquidation and whatever, and someone else is saying, well,...

...this ordering is good for me, and then combine them. So what it seems likely that we'll see on there too, is probably the exact same thing, where we have a multi tiered layer of marketing or like different mem pools that are then being aggregated into a final selection or a final sequence or like almost application specific, because man a, depending on a, depending upon the efficiency of bridges and latency of bridges across the myriad of networks that's going to exist in the future, being on the same network will have its benefits, and so you'll start to see kind of aggregates of applications that need synergy or composability, whatever buzz word you want to use for that, being on the same network, which then makes of a given pool of interacting transactions together, which then creates a market for how do we best order these these transactions? And that's where the that's where like that equivalent of flash pots may exist, of like the market for ordering these things right, exactly right, and that's very similar to what we see on l layer one. Right, that's really just as taking those concepts and matting them on to layer two or instead of mining, you have the sequencing and the mechanism that chooses the sequencer and then instead of flash pots, you have, you know, optimistic flash pots or whatever you so you're still playing with the idea of like what's the best way? If you want to keep this as a public good as best you can, then the fairness of that market on who gets to order is going to need to be pretty damn fair. And since you're not, you're not relying upon kind of arguably a fair distribution of energy across the globe to do this, you're you're using whoever we choose as a leader and whatever protocol besides that right and say like so the answers like we don't know yet, but we're actually looking into it. Yeah, I mean I think on the topic of like order fairness, like this is still something that's being defined and we'll see what for sure, and I think it's kind of it's almost beneficial that a Mev and it's kind of coming to the scene so hard and fast as beneficial because had you choosen something that was naive to this early, could have had a severe impact act on the game ability of your system. Yeah, that's that's most definitely true, and I'm unfortunately, I mean that has it to say that we'll see it play out and then it'll be clear what fair oring protocol that seems ordering is going to be an issue regardless where we go, like right and and another note is that so long as there are different definitions of a fair ordering protocol and there are then different chains that implement those definitions, then you're going to have envy, the arbitration between them, and that's what I kind of thinking. It seems to be here to stay. It seems to be so I see. So it's this technology. So you have you have two mechanisms here, kind of you have the the the technical process of rolling up transactions and watching the fairness of those things get put into the layer one and the and the like minimum viable implications of what it means to exit that system, and that's kind of like what like from from plasma to here. That was. That's what that's you're trying to build. How do we aggregate transactions in such a way that we allow people, allow more people, to execute things off chain and then embed those things into a layer one chain, while still allowing people to exit when malfeasance happens, right. Right. Then you have how the hell do we choose that person to order things and put it on chain? And those are two completely separate, different mechanisms. Yeah, and so what you're saying is that the technological one of how do we aggregate things and throw them into a chain? It is will go call it more straightforward. And by allowing people to come up with the different iterations of fairness on how you order, you have multiple layer two's vying for the markets in which people think is fair. Yeah, I think we'll see something that. That's that's about right. I think the question of fair ordering to me is still like completely unknowne. Whether one even exists or not. I think it's really uprade subjective. But like that's what some markets are for, right, as like that, people use what they think is fair and you try to make it as fair as possible based on what your definition of good is for fair. Right. Right. And the interesting thing is even in those so and the interesting thing is, I don't know, there's like this, like I don't know, somebody had debated like, Oh, extract the Mev, prevent the MEV. Right, the answer is going to be both. It seems very unlikely that we'll just get rid of all of MeV. In fact, I think it's probably impossible, and at the point in which that's the case, we need to have an economic grasp on what that is. So that's why I still like in many ways our original design, even though we are, you know, updating and improving it to take these...

...things into account, of being able to sell that first order right to transaction ordering at that, you know, distribute that by whatever fair mechanism you desire. But the end of the day it's not going to be perfect and it's great to be able to extract that value and fund the Protocol Development with it. I want to add something here. Like what I find really, like I meb, is not good. But what I like about this is that there's kind of a transparency in that. Like I meb, is parallel to a lot of things that exist in the real world outside of blotching that aren't aren't beneficial, like there's literally just front running of book at book at High Frequency Trading. Yeah, really, I like Robin Hood APP is like you're basically just, you know, selling information to aggregators so that they can skim more from you than you probably would get if you were actually paying to use the yeah, I appreciate this. That this, at least as blockchain does with many things, is make that dynamic, transparent and visible and you can choose the Mev regime that works for you. There's IT. Here's something that I because of what you just said, is potentially a possibility that I said, this is a problem that I was not a problem, but something I was somewhat worried about, that I haven't got an answer for yet or any type of like indication that there was an answer. And that is okay. We're using these layer two's as a way to scale the number of people who can throw transactions through this. But, as we've seen through the what Andrea Sinse Anopol those calls graceful scaling of the Internet, the more you give people room to do things, the more they're going to do them. Like the moment you give people the space and the cost of opening up new use cases or more transactions and so on and so forth, it's going to get filled up real quick. And so what I saw or worried about, and the end, like the advent of layer two's and everyone using them, is that it's this, it's the same shit. It's going to happen, which you can end up with. It was just a another layer and we're not going to fix anything, or at least like what's to stop optimism or whatever? The layer to do your is from just becoming the one everyone uses, because everyone likes composability and it may seem that like this this like concept of fairness and how you choose it at the sequence, your light layer ends up becoming a way in which you have multiple networks competing with each other, as opposed to just one being the only thing everyone uses and you just have their one being this giant roll up data deposit and not anything else. There's like there's no like, there's no reason to use it other than throwing data on the one layer two, and that doesn't seem to be like. This gives you at least like some game to play. That's like a a reasonable game to play. That's interesting. Yeah, that's a that's an interesting hypothesis. Side it definitely is the case to some degree that will see that playing out. I think the difficult thing that they brought up cries just to what extent is the are the network effects going to be fundamental and our suspicion is that there are strong network effects and having synchronous, like bidirectional communication that has, you know, atomic property of it. The flash a long where it all happens are at all dozen and that you know that that's really, really powerful, and that's why we really focused on on horizontal scaling, because you're totally right that you basically have to look, you have to try to look ahead at the bottom necks and and like preempt them, and it seems likely that it may just be the case that the bottom like that we run into, and I'll do is exactly when we saw an l one right, which is just the it's just the throughput. So it's really important to work on those things in the right order. Well, I it's it may solve a type of Mev case or problem that we currently see, but I guess that the question here is, is it possible for a once, once a layer, layer two, say what's the sequencer, throws all the data onto layer one? There's no way to change that, there's no way to mess with that, there's no there's no market there. So, like reasonably speaking, if say all of the applications decentralize applications, that currently participate in this in the in Mev market and we were going to test mostly Dex has things like this. Right, move to layer two. There is no layer one mev market anymore correct because it's all happening on layer two and you can't remember those transactions once the secrets of throws them in. It depends on the construction and I don't know, I think that's a good classication. Yeah, it's it's Inter right call on a tread right. Try. Sure. Well, I think like this is a weird edge case. There is like our fraud proving. So let's say that somebody committed fraud on on optimistic etherium. There's a lot of...

MeV around that process. The idea is that so if you could complete the fraud proof before someone else, then you get the pay out. The ideas that that is now that's a good work you want. I mean you want you want people fighting to prove fraud as best as possible such that there is no fraud. I don't and not sensoring that fraud process. That's all. That's a will know what if they're fighting, I think like if you're fighting too, I haven't thought about this a lot before, but if you have people really fighting to perform that fraud proof and bidding for it. Then you have miners who are have an incentive to include it and it decreases their potential incentive to to exclude that fraud proof from happening. Fingers crossed. I guess we'll see. Like yeah, I mean where I was going with that is it. If you have a bunch of different markets for Mev, then you've solved a lot of the current issues because it just moves it to a layer to where the concept of fairness is change. It's not like it gives you a an ability of for the layer one to just be basically a deposit box of aggregation machines, which is really what it should be in a lot of ways to made for various privacy reasons, if like scaling reasons, so on and so forth. And I've said, other networks decide what activity is happening. Here's a here's or here's a reasonably secure way to put it in, verify it and exit to if you need to. And like first case scenario, this is a really interesting pattern we've seen where something well, I guess I note it is the fact that these layer two constructions are built on a permissionless you know, smart contract layer actually isn't in a of itself, an extremely, extremely powerful thing. There are often times when we'll talk about ideas and they really are applicable to layer one, but there's just no chance of them being tried out on layer one because it's this, you know, incredibly important, hard to fork and important to not screw up protocol that is at the base of everything. But once you're at the point where you can just deploy a new clone of your thing with some new tweaks and some new properties or trying out some new research, it's super, super exciting. So I definitely think that we'll see. Layer two is allowing us to experiment a lot more. YEA, a part of that is it's not worth it. It's not a bit. It's like that. The time and money required to do that thing are very much not worth it. Until you get to a certain level of now it is worth it and as you start to scale into layer two's and there's more people doing smaller, faster transactions and a lot more of them sussing out what it means to be worth it to take it off that chain. In the case of Malfeasans of this very word in case of something going wrong. is going to be something that needs to be cleared up because, like for every implementation of layer two, there are if I want to say this didn't happen the way it should have happened and I want to file the complaint and take my money elsewhere, there's implications of that. There's consequences of that. It may take it's going to take time and money to do it and, depending upon how that thing is constructed, I might not be able to make that complaint. There's going to be a buffer of is it worth it for me to make this complaint? If I'm buying a donut and something goes wrong and it takes me a week and twenty dollars to get to the main chain to get that money back, I'm not going to do it. HMM. Have has that been thought about inside of like at least like the optimistic implementation of this? Yeah, it's a very interesting question and I think this all returns to the value of being within a synchronous state space. So what's fascinating is that this actually might be an argument for you to want to be on a larger roll up because then the because basically the more eyes and the more activity that is on your chain, the hopefully the more likelyness that someone else will take care of that for you, some big power, you spea even that is like what is the cost of taken care of it? Not Watching to make sure it can be taken care of, or finding it? It's what's the cost of doing it? Is it even worth doing? Oh, I see. So basically what happens if you're switching cost is higher than the value that you have. Is that the idea? Yeah, yeah, that's that's not a that's not a flat feet always. Well, I guess in most implementations, maybe maybe, as your knowledge one, they might be flat fee. But for I think for my intuition says that's going to be dynamic over time and I don't know how to I don't know how to measure it. Yeah, the these are tough things to measure. Hopefully, what we'd are seeing now is innovation on the bridging side in terms of...

...things like efficiency. So really a simple example of this is if you have a bunch of people that want to get out, they cannot basically all aggregate their exits or withdrawals together, and that can go through the main chain as one exit and then they get reimbursed on the other side to wherever they exited too. So hopefully that which do make it not linear in l one cost of the very least. But these are very good open questions. We it may be the case that's switching costs are prohibably high for some things, totally possible and that that's that's reasonable for networks right. Like so like, if you get embedded or ossified into a given way of things and and you understand the exit costs, then that's part of like the entrance fee of joining that community. You know what it's worth and you deal with that, you deal with you deal with that buffer and you only exit when it gets to the point of being worth your while or time and value. Something that I was worried about but beforehand that I haven't read enough or understood enough to to to fix in my head, is like, if someone exits do to do to fraud, does that ruin the chain? Great question, great question. So the answer is no, which is very cool, good, and that's basically the words that I'll throw out are subjective finality. So an important, very important thing about arms limitation is that you can achieve subjective finality basically at the same rate as l one does. So what does that mean? Basically, what it means is, yes, you there's a week long period wherest things can be disputed, but if you look at the chain and you see the correct proposal that you know could not be disputed and you see it immediately, then you know that the state is progressed correctly. You won't the LNE won't be able to acknowledge it because he needs to keep the door open for dispute, you know, for those next six days, but you know immediately. So that's that's the first one. Then the second. Do that in parallel. Yes, and the other component is that in the event that there is a challenge submitted that is a successful challenge, you roll back that second thing, that the notes were watching the state proposals without rolling back the transactions. So when someone submits the successful challenge, what that does is it says what was challenge was the result of these transactions. We're still going to keep them applied, but someone's got to come in and put the right result of those transactions because that one just got challenged. So you're directly changing the state, not rolling back to actual execution of the transactions. Yeah, you change the proposed state, you do not change the transactions. Okay. So if I word for word, like aality on those transactions, right, so to think about this appropriately. So, like, I guess we'd ever covered this, but if someone would submit a dispute, there's a week long period in which that think it's settled. Right, okay, so I submitted disputes and they went wrong. That separates from everything else. So you have this pool of things that are disputed in a pool of things that is disputed. So the way it works in our system actually is yeah, John, John many, I want to try to bridge the gap. Sweet so, if you like the like Ben said that the finality is like basically the same as what you get on Lne. So the idea is that if you're if you are watching Lne, you can read the transactions that have been committed to the chain on alone. If l one gets reorganized, those transactions might get get reorganized, but to the extent that you trust that a block is finalized on lne with some layer two transactions within them, you can trust that those transactions are never going to change. And then any observer watching the chain can play those transactions and calculate the actual state route. So so you can have a situation where perhaps the the commitment to the result of those transactions is dishonest, but everybody else off chain or reading the order of transactions can calculate for themselves and they should come to agreement about what the corrects, and so they all know the truth and can more or less go about their business. And that even like that transaction chain can continue to be extended during a fraud proof. So it's like you all, you all have a shared vision of the world, that is you can calculate off chain and trust each other based on and continue to transact over even though it's not like finalized in state and is still subjected to subject to potential challenges, but you can know that it's not going that challenge...

...is not going to succeed in rolling back those transactions. Then it just requires a little bit additional work. So like that, that's assuming that what I'm doing. So like I find if I am watching the state route on l one, I'm watching to see like, oh, there's no disputes, all things are moving forward. So any transaction on the on the network is undisputed. I can I just assume it's. It has the same confirmation, like the same finality as layer one. Yes, but even stronger. The idea is that even before it's finalized, after by that seven day challenge period passing, if you are honestly computing the outcome of those transactions, you don't have to worry about it. You just like you can see that the balance of eath does belong. Know, I can see o's God. My question is like to me, are are you expecting people to just watch? I'm trying to word this approperly because it's in my head but garbled up a little bit. If I'm trying to make sure that everything on up the optimism network is is valid, am I watching the state route or I just watching the transactions that I give a shit about because, like what you're saying is that, yeah, I have I have full transparency. I can view all the transactions, so if something gets disputed somewhere in the network, it doesn't really affect me unless it's my transaction. That right, MMM a little bit. Maybe the clarifying bit of it is that in the question is what happens in the event of it, of a successful challenge. Yeah, it's like Gus. It's separated because it like exacts. I'm assuming there's probably like, are you expecting there to be any challenges? Like how? What's the frequency of challenging? Are you expecting in there's in the system, and can that be dost? So, like there's basically always challenges happening, making people do extra work. Great questions. So yet two things that are one thing is, yes, there is a separation basically between a transaction chain and a state chain and there's a one to one correspondence between the two and the event of a dispute, is just rolling back the state chain so that the correct states can be proposed, so those transactions are still playing along. They never get taken off. That's one component. That's what we were talking about a little bit. got that. The other thing that you just got at is known as the verifier's dilemma, which is a fascinating, fascinating, like I can a problem and I think yet yet unsolved. In practice, we don't expect to see a lot of a lot of challenges. I mean while we're running the sequencer, which is doing the proposing right other other than, you know, bugs, we would have no reason to do to propose a you know, I'm mismatched state route in the future. How of this play out? It is more tricky to tell like this. This much is on. This was described as a problem before, before even what John got into, which is our finders fishing and Polka dot pink paper. Where the hell was right? Like right, exactly, exactly. Yeah, and at some point you have this. You know it's going to be a tough problem to solve. The good news that we see is that, and the reason that we're not like freaking out about it. I'm not running around right now with my hair on fire going, oh my God, we didn't think of this. Is gonna be no challenges, right? The cool thing is that people just seem to run nodes in Crypto, in the cryptosphere, like for reasons and and and it doesn't have to always be totally rational. Right, like there's people running l one full notes. They are not getting a block reward for doing that. I haven't. I don't get shit for it. Sorry. Sorry. So we've already seen this, like on our network, like we have infrastructure providers that are now running notes. Right, and at least at the point in which you're running a note, then it's a profitable action to go and submit your challenge because there's, you know, a bond at stake and so hopefully that that will cover our basis. Okay, well then, I mean, I don't much time I have you guys, for why don't we try to transition into what's next? When can we expect to actually get our hands on things? When can I deploy a contract? When is Madenet go? What's going on? Great question and great question. So, for so all you can do some deployment transactions today, when we have we've had our covent test not running now for for I think, happier happier at this point. Well, Gosh, almost a year at this point in some carnation. As far as what's coming up, our next our next thing that's coming up is a you know, swap release, which is coming very soon tem that's a trademarked date and and basically, basically that's going to be our next, next release. So that that will put out you know swap on the map. You'll be able to try that out. We'll be able to test the stability and load of our system can handle that kind of usage and everyone will go, go do that and then after that we're basically going to have a community release,...

...which is going to be the release that opens this up to arbitrary deployment and and stuff like that. So that's going to be the next thing out our road map, basically once we thought the eyes and cross the t's of getting the community ready for that and getting US ready to be able to support that kind of excitement, because even with an even with a lockdown system, people are always asking questions about how to integrate, and so we need to be able to do that scalary. But it's very excited. We're getting close. What help do you need? Like there someone is there? People you want, people you who do you want to reach out to you and ask you questions or help or do things with? Like what are you looking forward to or hoping someone does, such that like optors a moves forward nicely? Who? That is a great question. I will plug get developers. So I didn't get a chance to really show this, but one of the most exciting and unique things about our design for optimistic etherium is that we basically are able are able to implement all of our software as a fork to the goetherium code base to guess, and this is extremely, extremely useful for us to be able to use open source, the open source power of the world, because basically it means that if you are a gift developer, you are also a optimistic gift developer, and I'll two geft developer and those people are extremely, extremely awesome. So please, if you are out there, to take a break from your fearning shenanigans and come talk to us about how we're going to do on help to you and and about know for Real, that's where we buy the lot of these those that's where they are, and and come help us build the future so that there can be ten x transactions for for those from what and they expect to see different in the process of getting into your code based versus guests, whereas, where's the Gotchas that they're going to find? They may help them go from zero to useful. Honestly, if honestly like, what I would say is like we're constantly working to improve that answer. So really what I want to get them to do, to do is come and talk to us and say this is all the ways that you can reduce that, because they're definitely some ways we can reduce that. In general, the answer is basically twofold. Basically one you have to connect the node to the lone node, right. So it's the way that blocks are produced is no longer by, you know, verifying a proof of work, it's by reading from an LNEENODE and and putting those transactions into the ltwo node. So that's the one difference that you're going to see, is how our transactions ingested. And the other one has to do with something called the obm, optimistic virtual machine, which is this tweak to the evm that lets you run disputes on it after the fact. So Tho's going to be the main two sources of difference. Well, interesting place I find myself then. I usually ask what question should I have asked? That I didn't, but I have John here so he can ask it. It's you want me to ask the question that you should have asked. Yeah, you can play both roles. You're that again, I guess. I think. Ben What like maybe asking about did we talk? We talk enough though, public goods in this conversation and like what we think is the right like where the proceeds of minor extractable hmm, you would guys have directed. That's a good one. There you go, that's a good one. That's a great one. Yeah, so one of the most one of the most one of the things that we're most excited by right is to basically put our money to our mouths in terms of this public had stuff. So we have this public benefit Corp, Corp, not court and and it's very, very, very important to us. Not Because, well, because of our values most definitely, but I think even more fundmentally it's just we think that our success is completely predominant on this. Isn't is the open source community taking this over, and one of the great things that we're seeing right now is a lot of innovation and public its funding right. Like, I don't know if you guys are watching all the Getcoin, you know, grants arounds that have come out over the past year, year, two years now. But one of the things that we identified, certainly when we were a research or, you know, just putting out open source, you know, papers and code, was that you have the problem of funding. Where does the funding problem come from? So something that I could have that I maybe could have should have highlighted more there. So thank you, John For eraising it is that when you do this auction of Mev of minor checkable value. The reason that we're most excited about that is because that is a revenue source to put back into funding the protocol. So right now I want you ask me what I want to happen. Who should come talk to me? I want get debs to come talk to me right now. But what is going to have to happen in the future for us...

...to be successful is get get debs to just go start writing and contributing code and be rewarded by the protocol for contributing code to the protocol. So that's saying that we're extremely excited about. We have some post coming out very shortly about some of our thinking around what are the right mechanisms to be doing that by. So I'll leave that for the post, but it's the right callout, John. It's very, very important that we make this happy Rainbow Universe possible by funding people around the world contributing every pull request. I look forward to reading those blog posts. We're going to go to find those blog posts. In the previous image made twittercom slash optimism PBC should have the link to our medium. If an inscription. Is theerium optimism yet should be in the description or optimism? That I will definitely have it. Yeah, okay, Doctor, the both of those links are already in the description of this of this video, so go down below and check them out. What about you, guys? What do you what do you most excited like? What are you most excited about over the next year? Like where do you see this stuff going? What? What is your optimistic, sorry to use the word, view of what is serium looks like over the next year, because it could go in a lot of different directions and I don't know what most people think. John, you first shit, and yeah, I mean broadly. I don't know how how well this ties into yeah, it does. So so I think, like the thing I think a lot about is is like, instead of saying cross chain, I'm starting to think in terms of like cross domain more and more, and a bit of almost like a melting away of the idea of layers. So I anticipate a future where the it's like you think less about like in my in the sense of like moving from l one to Lto and more just like moving from one roll up to another roll up to you know, like a cosmos zone or potentially what is poked up pair chain, like I and then and then also between a syrium and the theorem of shards, like I see, there's going to be some weird mental models and user experience flows and security challenges around these these different domains that are cropping up, and that's going to be really fun. You know, it's not fun. We've been, it's spending a good portion of our time trying to figure this out, its status, to try and make an interface that works in this world. Okay, so like there's a security guy that's amused by when things break. It's okay, that's a better way of put yeah, yeah, sorry, sorry for the people that actually have to use it and great, but that's my answer. I Agree Been Oh man, I'm going to sound like a total east show right now, but I am pretty excited for you. F You one hundred and fifty five nine. That's not like that's like a bag holder thing. That's a use it in our protocol thing, like yeah, I p one hundred and fifty five nine and having this access to this base fee, that will basically tell you, like what is the opportunity cost for a minor to sensor a transaction. I think is going to be super, super valuable for layer two protocols. From a just from a future information perspective, being able to predict a little better than you previously good. Yeah, and we're going to be able to do better as well on basically the economics of these challenges, because right now it's kind of hard to denominate a challenge which is a bond which is meant to recoup the gas cost of a challenge, because this is just like a hard problem. So I think we'll start to see things, well, even in the extreme, I think we'll see really cool synthetic assets being built around the base de Price. But yeah, this is the kind of this kind of like a boring technical answer, but being able to economically priced bonds better for roll ups is can be Super Dope. Write that down. I'm a put the quote for you. All right, guys, I appreciate it. Just hitting over an hour. I think we can go for another one if I really wanted to, but just want to have you back on and thanks. Come on show. Thanks having this is awesome. Thanks, Great.

In-Stream Audio Search

NEW

Search across all episodes within this podcast

Episodes (128)