Hashing It Out
Hashing It Out

Episode 48 · 3 years ago

Hashing It Out #48 - MakerDAO pt. 2 - Nik Kunkel

ABOUT THIS EPISODE

MakerDAO has built a token that is reasonably stable with the USD. We have the great pleasure of interviewing Nik Kunkel, Head of Backend Services at Maker. In this Part 2 of 2 on Maker, we talk about the technical design behind developing Maker. Please listen to Part 1 of 2 (previous episode, #47) before this, as it's essential for understanding the principles guiding the design decisions. This is compelling work, and truly amazing how well it's standing against the throes of the fluctuating markets.

Links

https://makerdao.com/en/

https://www.reddit.com/r/MakerDAO/

https://twitter.com/makerdao

And 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 and the problems they face along the way. Come listen and learn from the best in the business so you can join their ranks. Welcome back, everybody. So forty eight, part two of the maker first part we have that counckle from the lead of back in services. I am Dr Cory Petty, and My cohost Calling Cuche, Aka Craig right, or the opposite. I don't know how you want to say that. No, no, I'm road. You're here now. Okay, sweet I'm just all the people all. Put me on all the white papers, put me on all the scams, put me on everything. I'm going to be gone when we get sued. You technlo next, and then, yeah, just this keeps ball rolling. A man, I'm just speak whoever the hell I want to be. So if you're listening to this and you did not listen to the last episode, stop, go back and listen to that episode, even if you think you know how maker works quite well, I'd still like to. They'll like. You should probably listen to it to see if we tell you anything you thought you knew but didn't know, or we got something wrong. I assume we didn't, because next with us he would have told us. But this episode, last episode, was basically giving you a quality overview of I'd say the maker die ecosystem, how it works, but the how the value flows, kind of the INS and outs, and some and some, I guess, if you frequently ask questions that a lot of people may have around the system. This episode we're going to try and dive into the engineering side of it as to how it works technically. We have quite a few questions. Are and what's going to see how the conversation goes. So welcome back, Nick. Appreciates you doing this. Yeah, let's let's start. I think I don't know where you're starting. Many oracles. Yeah, Nick, what's a smart contract? Okay, with our audience, those that already so. Yeah, oracles. What are they? What are you guys dealing with them? How do they work? You asked the right guys. So me and Marianna County, also a maker, we're the oracles guys. So our oracles were actually mote like used by everyone in the space. I don't know, I think this's like a little known like fact. I actually use it. So I I've made a donation APP that does that kind of automatically converts how much people have donated to our podcast in the bitcoin network podcast and not to to find the price of ethuf dollars. I use. I make use of yells to work all. Yeah, I mean that's really common. So, like all of defi's pretty much using it right now, like set compound, dydx, Darmagnosis, augur, all of these guys. That's a part. Why can we trust it? So it's essentially it's Federated Trust. So we talked the last episode how, I'm care holders make all of the decisions, while they also decide who they want to be an oracle. So who is allowed to be pushing prices? So right now we have fifteen oracles that are all kind of pushing prices for for Atherium, and what the smart contract effectively does is it takes the medium of those fifteen prices and that's the price. That is a that's used by the system. So, funnily enough, so I've been what I've been working on the past year. Actually is our kind of next Gen Oracle architecture that we're going to be releasing alongside multicolateral dyed, and I think it has I think it has a lot of advantages to the current model. So in the current model, everyone who is running an oracle client, so all of those fifteen people are organizations. They have a smart contract that's that's owned by them. That's kind of like a cash so it's like where they kind of push their prices to and then they kind of poke our medianize our smart contract, which, you know, reads from all of them and takes the medium. That's kind of that's kind of an inefficient model because it means if you have an oracles, you know, if you're doing n price updates and transactions, and so when you think about must like collateral die and you think, well, you know, we're going to have, you know, potentially hundreds, if...

...not thousands of assets instead of just one with eeth like we like we have right now. We needed to come up with something that is much, much more scalable. So effectively, what we did is anything that you could put off chain in some kind of verifiable manner. We did so there's a very, very little left to do on chain. So so what we do right? So what we're doing in the next gen is that all of these oracles, they're now and not pushing their prices on chain anymore. They're not pushing a transaction. Instead, what they do is they sign a message, and this message is it has the asset parent name, has the price and has a timestamp and they sign this message with their theory and private key. And so now you've basically created this this signature. Right, so you have the price, you have the time stamp and you have the signature that verifies that you know this came from from a valid oracle, and you put all of these oracles in this peer to peer gossip network. So the particular one we use is called scuttle bont, but it's it functions very similar to to other gossip networks and you know, for for your users, maybe moren't familiar what that is. You can kind of think of it like twitter. So if I'm following Corey and Corey is following calling, and calling kind of post something. Since Corey is following calling, he's and he sees that and he's gonna think of it like retweeting it. Now Corey has posted collins message and now, since I'm connected to cory, I see the message that Corey posted from calling. And so you can see how this kind of propagates through the network very, very quickly, through these kind of peer to peer connections. And so effectively, what you have now is you have this peer to peer network of all of these gossip price messages and you can have these third party entities coming and so in our in the maker ecosystem, we have this concept called a keeper, and what a keeper is is it's really just just any it's like a bought that you run that it makes you money, right, it does some useful service for the system and it earns some type of patrol. And so what happens here is that so, for example, core you could be running a keeper and effectively you would become a member of this peer to peer gossip network and you would see all these price messages and you would collect a bunch of them and put them into one single transaction that you then push on chain and you get recompensated, right, your gas cost plus like a little bit extra feet. And so that's how you incentivize these third parties to push these transactions. And so you'll find now is that instead of having an oracles and and transactions, you can have an oracles one transaction and that scales much, much, much better. So and what we've managed to do there is actually produce our costs by about ninety eight percent. What are you to do about maybe like spam protection and griefing this type of that work, because, say, multiple people can be exacted? Saw The keeper, I imagine, would aggregate and take the media of all of the different oracles that he's receiving based on some given asset and then publish the medium of whatever he has for a given asset based on all the RBS network. I imagine that gets validated only by the signature of appropriate people who can be oracles right, because they're just they're they're basically setting a message and passing it on. Is there a way to like perform a clips attacks of this type of network to try and corral the choral and given price based on collusion, is there a way to maybe spay on the network so you can kind of Dedocet and stub my Champe or form, because I could work for work for status. We've definitely worried about these types of things quite a bit. And so when you when you change from a federated network where you have a lot of redundancy at it's the cost of efficiency and try and turn everything off by lowering those costs, you may be potentially introduce attack vectors or someone can can can sway thix, have you like? Have you thought about these things and know so, what have you thought about? Yeah, I mean, of course those are all very valid points and we couldn't. We can definitely get into that if you want to do that. So the effectively, what the Game Theory comes down to is can you get fifty one percent of the oracles right to colude? Right? So it's very similar kind of in terms of like a blockchain Hash power right it. Can you get one percent of the Hash Power? Can you get fifty one percent of the oracles to to colude? And so in...

...terms of spam, for example, there are kind of built in kind of regulatory kind of behaviors that each of the oracles is running. So if one oracle is behave kind of maliciously or trying to spam it or de Dost like, it, can just cut it out for example, I can put it on like a thirty minute time out, you know, and like you know, if after thirty minutes it's it says hey, I'll be your friend again, and it's still, you know, spamming it, like Nope, you go right back to the timeout corner. You also have, well, this I'm still kind of working on, but I want to kind of make into a staking model, right where effectively, oracles put up steak and then have some kind of challenge response mechanism. So if I'm an oracle and coreys moracle and Corey, you're putting out some what I think are malicious prices, I can send you a challenge message and you will recognize that challenge message and you can either ignore it, which case you know your steak will get slashed after some amount of time, or you can respond to it. And you can respond to it and say, oh well, I actually, you know, I checked my sources, my price sources, again and I you know, I stand by this message that I that I put out. And now the the challenger, basically has to put up their steak. Right. So now it's say both the both of you have have committed, versus, if the you get challenged and Corey, you go back and like, Oh, you know, maybe the price source I was using, the API, screwed up and gave me a bad price. Actually, what I meant to say is this is the right price. Right, then the the challenge has been responded to and everything is fine. Now there is you can kind of put reputation at plant here too, right. That's what I was especially if you have, I guess, a contingent upon this, and this is this is the difficulty of reputation systems. Is it? Is it a anonymous oracle system, or is it a federated Oracle System where you actually have like a set number of people who are capable of being oracles and then they do the network? That makes it a lot easier in terms of during reputation. So it's actually both. Okay, so we kind of differentiate this by by two terms. Are called them dark oracles and we call them light oracles. Dark oracles are individuals and it's imperative that they remain anonymous, right, because if you are an individual you can be targeted, you can build blackmailed, you can be coerced. It's just much better if you're anonymous. And and also that means that the Oracles, it's much more difficult for them to coordinate include among each other. For Light Oracles, these are public facing oracles and these are not usually individuals, are usually organizations like institutions. So, for example, you could have compound be running an Oracle, you could have zero x running an oracle. You could have, say, a coin based run Oracle. Right. And and these these institutions, right, they have a lot of users, right, they have a lot of credibility. If they were to act maliciously, you know there is a legal entity that you can sue. Right there there the instent, the Game Theory Dynamics like play out that they can actually be public facing, right. It's actually better for them to be public facing because of this reputation that they have at stake. And so another thing you can kind of play around with this is, remember we talked about like a fifty one percent attack, you know, being the thing that breaks the camels back. Well, I mean, if you can start waiting these things too, like you can say I want to wait light oracles more than dark oracles by some waiting, right. Or you can break it down to the individual. You can say this Oracle has been you know, around for four years and there's always been, you know, Ninety nine point nine, ninety nine percent up time and the like. You know, I want to trust this, why? A little more than the oracle that just got added six months ago that, you know, seems to be running well. But you know, who knows, and so so maybe we can move on to something as I don't I don't want to take up the whole time talking about oracles. Well, okay, so oracles are pretty much so what are the categories of of use cases for the oracles right now in your system? Literally just you have price identifiers, you have auctions. I guess would all be oracleized. That you wouldn't want to run that directly on chain, would you? Like I think meant, but like last episode, you mentioned that they're on chain. But like, is there a possibility for having some truth mechanisms that exist for some of these on chain calculations which might save some transaction fees that in the future you might want to velp? Or so the everything is on...

...chain by design, and that's to make it decentralized to the point where even if maker as an organization were to disappear, the system would completely work the same way that it works right now, so that when we're designing something, that's that's the most important piece. The second most important piece, then, is, okay, the the gas optimization. How do you structure this right so the amount of value leached by by fees right is minimized? What we use the ORACLES for? The only thing we use it for is to know when a loan can be liquidated. That is that is the only piece, okay, okay, and and there's really no cheap way of doing that on chain. I would take it because, first off, you're trying to go with actual timeframes, like, and you can't guarantee that or like. What's the reasoning behind needing that off chain, the having the Oracle stuff work off chain right now? Yeah, the liquidation side of things, it's outside information. That's the basic ideas. Like, that's as any oracle requires outside information and because you're pecking to a dollar, that's whatever you end up pegging to, you're going to need some type of reporting of that, of that thing, and that can't be done without okay, so that's part of the liquidation. Take. I got it. Okay. So when I was thinking liquidation, I was think of the actual act of like calling liquidation on these things is like you can just submit the truth value to the blockchain and and that has an effect on the entire smart contract ecosystem. Yeah, that's what I was kind of like thinkings you actually individually would call liquidation. You're not doing that, it's just submitting the value. We don't call the liquidation, right, we just supply the price value. So when someone else, like a third party, remember we were talking about this keeper, this like self interested third party actor, they're the ones that are looking for which loans can be liquidated, they right, right, price finally drop low enough to liquidated. Now they push your transaction saying I want to liquidate that one. I got you, and for me that was kind of like the Oh so, wait a minute, they want to they want to liquidate that one. I got you, okay, in that communication is wait a minute. So actually, now I'm confused. So they say they want to liquidate that one. The the the maker holders. Who's that, charrectors? Is that not the maker holders? Anyone? So the keeper is just a plot that anyone can run. You can run it or can run it, and the reason why someone would run a liquidation keeper is all it is, because, you know, you get to be the first person to you get to guarantee that you were the one that buys that collateral. Right. So I want to liquidate that loan because the prices of youth has fallen far enough that it's under collateralized now, and I want to buy that eth. You know, a three percent discount. That is. That is and you can profit off about three percent discount by selling that either on you to swap, by selling that eth on Kiber or, you know, whatever decentralized e change you want. So you've basically tried to find a way to get just anyone to systematically watch for the types of things that need to be called on chain and you pay them out by giving them a premium on buying it. You leave people profit opportunities to make them do things that are useful for the system, cautch. So okay. So the main component, I guess, of makers the discard contract system that is in chain and then the oracle system. What else are we missing? Is there anything aside from that that we should be talking about? I I mean I know that sense reduction. Reduction is a smart contract system. Yeah, that's but that does a two major components, like literally, that's the kind of the point of what I'm trying to make here is that, from an architectural standpoint, to develop a large scale financial transaction system, if we get the tooling down and we get the processes down and we get the best practices down, building something like this is a lot simpler than building visa rate and and it's a lot simpler than building a BANQUETFRA structure. And so I didn't mean to say that as in like you didn't put an amazing work, although it could sound like that. What I was actually saying is this is what we have and this is all it takes to develop something in the only reason that we can do this is because of this concept of trust, this decentralized trust mechanism, which is the Blockchain, which is the foundation for the entire thing that you're building. So I just I just want to point that out. I thought that was interesting that you have this oracle system, which is actually kind of, you know, an ecosystem of Federated Truth, and then you have the blockchain, which is the absolute truth, and these...

...this automated logic, which a base of predefined set of rules and people can interact with and it the Yada, Yada, yea. But the problem that people keept running into, and this is Corey's area of expertise, I'm sure he's going to have a ton of questions in this area, is the matter of security. So as far as a smart contract security goes, how did you verify that these things operate as as expected in as many circumstances as you can predict? Okay, so your you're absolutely right. Security is really the the most important thing. Um Right, we have about two percent of the entire theorium supply locked up in our smart contracts right now. It's just above two million either, which is you can think of that as like a pretty significant systemic risk, right. So how do we protect that? So we have been working really closely that, like most of the top audit firms in the space. So trailer bits is someone that we work extremely closely with and they're not just a one of the top, or if not the top, CRYPTOA cyber security from we're actually like one of the top cyber security from s period in the world. We've also like, I think, in totally we've had like five or six audits. So the the white hat hacker group, those glorious anonymous individuals. They have audited the system as well. But personally I don't think that's enough. And so what we've done with multicolateral die is we've taken the next step and we've done something that no one's done before, and so we've actually formally verified all of the multicolateral die contracts, every single piece. And you know, I'm sure your users are super knowledgeable know what that means, but for the couple who don't, will what formal verification and means is that you can mathematically prove, you can like basically algebraically prove that the system can never do this, it can never do that. You know, given this, this can never happen. So, for example, you could say the system, you know, should never be able to print more die than the sum of the debt ceilings of of all the collateral times. You know, the system, the user, should never be able to mint more die than their collateral is worth the use. No one should ever be able to withdraw die, except out of this one function where when you're paying back her loan and you get to unlock your diet. You should never be able to withdraw more you know eith than you put in. It's literally just say by you start very, very simple and build like these little primitive truths, and then you can start using the like Lego blox and stacking these truce on top of each other to prove very, very kind of complex relationships. Now, so one awesome. I'm happy that that's the role that you went. I've have some questions on how you like, what software you do that and you did it, and which remain a very significant point about that is that, particularly in the theory him, you can only form of formally verify the truths that you put into the system in check. We're never going to get a system that's basically like all possible scenarios are set within the context of the evm right, and so the quality of her former vocation is basically at the limit of the quality of the rules you put into check in a lot of ways. Who did that? Like? What did you use? Like Kvm? Did you use manticore? What system did you use to do this formal verification? We actually do both of those. So KVM is basically we use the K framework to to kind of write our proofs, and this is actually some of the same kind of use, a lot of the same tooling that trailer bits use this internally. So you mentioned mental core, right, mentalcor being one of them as well, kind of like puzzing and stuff. So you know, we've had, I think at this point we've had like two or three rounds of all this with trailer bits, and the last one, hope I'm a want to say this publicly, but the last one came back and you know, these things are usually like twenty pages thick, right, and they're like, you know, they list like Oh, here are the high severity issues, the medium severity issues, near the low severity and by this point we'd already formally verify them and they just came back like look like, we're really struggling to write anything here. The the report was like three pages long. That's that's that's it's almost like scary she gets something back...

...like that, to be honest, because it's like okay, maybe it's so complicated that even auditors are having trouble with it, but like it's definitely one of those situations. Were like good job on the engineering speck was good. We're having trouble finding issues to discuss with you because it was done properly, especially when you do it multiple rounds. If you want to have a comparison for what one of these reports looks like when it's not good, go to the trail of bits, Github and you can, I hate to be a mean person, called them out, but look at the audit they did for basis, you know, previously called base coin, and they also have a repo called not so smart contracts within within their goodub. That is a quality amount of like good things that will come out into report. Yeah, so that's that's really awesome and I guess a part the next situation of those I ke've audited all the all the things that you're going to be doing for Multi Claro die and things are rolling out. The next question, like the next part of that, is one. How are you going to monitor these things so that they're operating the way they should once they're deployed and to say something changes in the upgrade path of a theme? What is the upgrade plan for these contracts that have you have you have you thought about those? So the M K, our governance, has the ability to basically we read the Bloy the system right. They can write by this system has ownership over the die token and the MK, or so it can write, so they can mint and destroy them. And if I'm care holders want to upgrade the system right, they can do a governance vote and then there's a basically they give the new system gets deployed and they give ownership of those tokens right to that new system right. They give ownership of the governance mot they make the governance module control that that new system. So there's a very clear path here for for upgradability. So say, you know, eat two is is starting to be built out, right, they can. They could have a new system or they can have a system in parallel. An interesting point here is actually when multiclateral die comes out, the current die, which we kind of have re Christen single collateral die, that system will run in parallel for about four to six months. Those two just to give people time to migrate over and transition over in their own time, when they feel comfortable, you know, when they want to, when they think market conditions are favorable for them to pay payback and close their positions. We definitely don't want to be the arbiters of saying like Hey, you know, force migration now, that's that's not how this thing works. Right we we deploy it and we give the community the option of using the new system and we give them ample avenues to transition over. So if you want to see if you have a CEP open right now and you want to transition that CD the multicolateral die, will have a mechanism to do that. If you have single collateral dye and you want to upgrade it to multicolateral die, will have avenues of doing that. So you know, we for example, we talked about the oracles earlier. You know all of defies using makers oracles right now as the price source for Ethus d you know, we don't want to screw those guys over and we you know they have to plan from an operationous point of view how they're going to migrate to to the new oracles and they need time to do that and we understand that right like we're not trying to hurt anyone here. So we kind of we understand that all of these systems are are interlinked, which is great, but the weakness of composability is right, that there is a there are centralized points of failure and you need to behave responsibly when you are at that base layer infrastructure and providing it for the entire ecosystem. Yeah, man, that's a lot. So, yeah, I mean like I'm looking through your ducks as you're talking. By the way, some of the stuff I got, I'm actually looking for you smart contracts so I can look over them and see what design principles you like, how you actually building them and stuff like that. I want to see this. You know, formal verification methodology that you were talking about. You have a very large repo, though, so I os them. Going through a lot of a lot of stuff is not going to work right now. The DMC code is and is it DSS d? Make your okay, yess, nice stability system. That's the multi colateral Nye Code. Okay, and you'll find that the contracts or not that big, like really you've made a lot of effort to really hair them down and make things as simple as as they possibly can be. Gotcha, okay, and that...

...is that is good for multitude of reasons. It also means if people could just throw up their own version of this and just play with your model and have compatible versions of maker just running out there if they really wanted to, although I wouldn't see a huge need for that if this is working. Okay, so I the real the question I was about to ask and then before I realize, is that I'm curious about die digs and how you how you foresee people building decentilized applications on top of this, and you know some of the some of the design personals. But on that I understand that you are the back end guy, but I'm pretty sure they t testa of that. You have to also do some front and work to and so I'm kind of curious. You know what is what is your vision? What is the vision with Die Js? and Are you guys expecting people to maintain the point of entry for people interacting with your stuff, or do you see people building alternatives to this? So both right. So we we want to have kind of a vibrant ecosystem built up around maker and around die, and that kind of starts with or knowing where we want to be positioned and leaving everything else to kind of everyone else. So we want to be on the base layer right where we just generate die and get die in circulation to make die useful. We don't want to like, you know, soak up all of the like the opportunities that you can build on top of die, because it's much, much stronger when when it's a platform, when it's a vibrant ecosystem. So we want to leave those little pockets of value, therefore, for people to do. So what we found already is that there's a pretty big ecosystem building on top of maker and and they'll most of them are using diet JS that you mentioned earlier. So, for example, there has been all kinds of like alternative front ends that people have, people deployed. So there's a there's instant d AP, there's there send wire. Are Not not send? Why? Our whire? I think it's stable wire. Right. Well, whyre yeah, it looks awesome actually from a conceptual standpie. I don't know if it's like functional yet or working, but again, it's revertial payments. Yep, you basically bought the systemic enables. One of the things that kind of bothers a lot of people about this crypto stuff is that there's no way to actually like for herds of payment going wrong, which, you know, some people see this vage, but a sort of situations that might not be and one of the coolest things I've seen actually something called maker scan. It's speaker ether scan well, but for all things maker. Another really cool one is maker dot tools, so NKr dot tools, and you can literally see, like you can look up any CDP, you can see how much collateral as you can see how much debt it has, you can see the entire history of actions. Well, on this date and time, this person locked up so meth and then they borrowed this much die and then on this day they paid some back. Oh well, this is really pretty, really really cool. Yeah, it just the thing is like in blockchain we kind of take that for granted that there's these cool things, but take a step back for a second and think of maker. Like the maker kind of system as a bank. But this amount of transparency does not exist in the traditional banking world. If you will. Bank of America down the street and you're like hey, I'm thinking about taking out alone, but please let me look at your entire loan book first. I want to see every loan you've extended. I want to know every individual you've extended it to. I want to know the duration of the loan, the interest of alone, like before I make my decision, if it's like safe for me to take out alone with you. They tell you to get lost. Right. It's the amount of transparency that we have in this system. Like right now it seems very opaque to regulators and you know they're they're understandably fearful because they don't understand it yet. But once they realize how much war transparent this type of system is, then the traditional banking sector, I think they're going to fall in love with it, the most easily regulatable thing they've ever see. Now that you can always build. Try to throw this contact concept out as much as I possibly can, but you could build centralized services or even non transparent services on top of transparent services. You can't do the opposite. You can go from you could, you could, you can go from decentralized to centralized, and put's nice about that is you still have the basic on like reliability layer underneath. It gives you transparency, and rebuilding a fight the financial world on top of this is definitely doable, one of my opinion, like inevitable. But that doesn't hurt the AV the base layer in a lot of ways. In fact, that...

...it boldens it and throws a lot of use cases into it, because you have a massive funnel into the base layer. And but you've kind of built, if you're saying like kind of what we came back from yesterday, as you've built a decentralized credit system. In a lot of ways, what has been don't kind of throws me off. Full of it because, like I think the they this is credit system. I guess you're not like yes, you're putting up collateral, but you're not really like that's not credit, like credit is. Hey, I am known to be pretty good, so they're going to give me free money just because of my reputation. This is like you're putting out collateral. There's real risk involved in you being a bad actor. It's not like you know, like yeah, the other with a credit score like you be a bad actor, what happens? They might be able to garnish wages, they might not, and they'll be able to hurt your credit score. So you can't do that chronic jump again. But the risk is all on the person giving the money. The money is not being given to you by a third party. It's not like this is not the same thing as credit. If you it's closer to Dharma. But I don't even see acting like Dharma, because Dharma's kind of this peer to peer system, but peer to peer like. In fact, that might be a good distinction to make here, the Dharma Protocol versus what you guys are doing. It actually seems like that must work well together rather than being a competing thing, which but I originally kind of thought. So maybe that would help with that distinction that I'm trying to make here. Oh, I mean, we have a great relationship with with Dharma and and all the secondary learning protocols, right. Actually they I just want to I know you ask a different question. I was want to touch on this real quick. All of these secondary learning protocols actually hate help us stabilize die because effectively, if we offer, you know, some rate and you know someone else has a more attractive rate, right, they'll just go over there and borrow die from them instead of borrowing die from us, and that actually helps reinforce reinforce the pack, right. So when the thing is where we move rates, they pretty much have to move rates as well, though it's a symbiotic type of relationship, because any die that's being lent out at Dharma, someone is paying, you know, the whatever our stability fee is. You know on that die. You know every second. So you even though it's not that exact individual who is was paying that from a macro level, somebody is and that all actually tends to equilibrate out. Yeah, so that's awesome. You built literally an ecosystem, a platform that people can play with nicely and there's really no competitive advantage to doing anything wrong or right. It does I don't not seeing the more questions I have, the more they just seem like, yeah, why would anybody that? You know, I mean it's the best answer that you have. Or you know, it would just break the system or they would just lose money or it wouldn't. It wouldn't help anything. So I think that's interesting. But you know, there's always social attacks outside and reasonings beyond. Like if somebody one attack etherium, the fact that you hold two percent of that would be a pretty major attack if they found some way of like gaming the system to attack that. But, like most people, don't know. I don't really see somebody having that kind of money laying on to do that kind of joke. I think a better way to maybe yeah, there could be problems, but we're relatively naive people about it. Why don't we ask the back end guy, like what's difficult, like where do you see potential issues in the future? Where do I see potential issues in the future? Oh boy, or like what's maybe starting off with, like what's really, really hard to get right from an engineering perspective, so that, like, you have this quality mechanism, design, this equilibrium, and how do you know building it? That's what you're going to get. Yeah, what talks about maker man or so I think. I think that a flaw that we have with the currents die that is released right now is that the only lever we have to really stabilize the peg on a macro level, right, was this our ability to in to disincentivize people to create die, right by raising the interest rate, right, or incentivising them to create die by lowering the industry. So that is a lever on die supply. Well, we do not have right now is a lever on die demand. And so the way that we're addressing this, and most like clatter all die, is that we're going to add something called the die savings rate, and it's effectively are this lever on demand. And so what the die savings rate is is just like how you can go to your local bank and you can put some dollars in a savings account earn some interests. It's going to work the same way. So...

...you can lock up your die in a smart contract and per second it will it will basically accumulate interest. And so the where this interest is coming from is it's actually coming from the fees that people are paying for the loans. Right. So say so, say right now we have nineteen and a half percent interest when someone takes out a loan in Beth. That's pretty high. I think, well, great, yeah, well, sure it's, but it's pretty high. But what if you could instead say, well, why don't we like give people six percent interest annually just for holding die? Right now? People now there's going to be an increased demand for holding die. So this equilibrium equation has shifted right, because there's now increased demand for holding die. And now the system can actually withstain increased demand for leverage as well for creating new supply, because the demand equation has increased. So we can actually use the die savings rate then to lower the interest rate that we're charging borrowers, right, and so we can bring that nineteen and a half percent, say you know from you nineteen and a half percent interest, great, and zero to die holders, and bring it down to, say, ten percent interest rate on borrowers and six percent to die holders. And so effectively those die savings rates gets funded from from that stability. Feel right, and so you can. It's effectively a win win for for everybody, right. I mean, I'm care holders don't make as much money, but it allows the system to scale more so even in the long term, and care holders can can recruit those costs through scale. But for die holders and for borrowers. For borrowers they get lower race and for die holders, right, they get to earn interests, which is great, and I think the impact of that is is going to be felt, be felt kind of ecosystem wide when people realize what it means that you can earn six percent interest on die or four percent or you don't, whatever it ends up being, because think of all of these wallets right now that don't have a business model. Well, if all they do is just you put their users deposits and lock it up in this interest earning account and they can write and they can take it out at any time. Right now, well, let's have a business model. Think of exchanges that haven't listed die yet. Changes are you know, volume is overall down from where we were at two thousand and seventeen levels. Exchanges are trying to find more avenues for more revenue, right, and that's why they're doing this whole ieo thing now, because they've had to diversify and find new areas for revenue. Well, what if all of the Diet that their users deposit on their system can be earning them, you know, four or five, six percent annually? That's another income source. So now exchanges are incentivised to list die. Effectively, any application that is, you know, holding die in some way now has an extra income stream. How do you how do you see people adopting this? Do you feel like it's going to be institutions and applications that build on top of this system, or do you think? You think indusers are going to like, directly interact with this? I think users will directly interact with it, but it has to be in collaboration with the the ecosystem, right. So I don't see the end the die user in five years from now really interacting with like, you know, the maker front end or anything like that I see it as like someone built an APP in Argentina or something and they're using die in their back end. The person probably doesn't even know they're using blockchain, using CRYPTO, using die. They're just like, oh, I have money and they can go and their APP, you know, is saying, Hey, we're going to give you three percent interests or something like that, and the APP is making money because they're pocketing the difference. Right. I think that the die savings rate kind of incentimizes this whole ecosystem of participants to go out and kind of be the missionaries for for the use of die. And it doesn't have to be hey, here's this cool thing, die, use die. It can literally be abstracted away to the point where you don't need to know you're using blotching, you don't know you need to use crypto. One really interesting use case I'm super excited about is kind of tokenized securities. So we're working with a partner called trade shift and they're one of the largest trade finance kind of companies in the world and what we're doing a pilot project with for tokenizing invoices. That's what I was kind of kind of gets to you I guess some of...

...this I wouldn't know yet, like token I thing, like that's basically an n FFT. But yeah, sotally, yeah, it has to have a valuation. It's like how do I how do I take collapse? How do I take die out on my crypto Kitty's basically the same problem. Yeah, well, like, I want to hear all about this. Go, okay. So in in the trade finance world, right. So, say you have Walmart and you have some small kind of toys manufacturing. It's, you know, Christmas time is coming up and Walmart wants to make a big order. And you know the Christmas toys manufacturer. All right, they have some cash, they manufacture some toys, they ship them the Walmart, and won't they issue Walmart like an invoice, right and say like pay, okay, you know, you owe us two hundredzero. Walmart doesn't pay that invoice immediately. They you know, they'll usually pay it ninety days from the day that it was issue. And this is a problem. This is a problem because supply companies are very cash flow dependent businesses. Right. So this toy factory, right, you know, maybe it can't handle any large orders for a while because it doesn't have enough capital. So what they end up having to do a lot of the time is end up selling their invoices for, you know, say seventy five cents on the dollar. You know, sometimes even like fifty cents on the dollar, depending how desperate they are. Actually there's a pretty predatory strategy that Walmart and Amazon sometimes employee where they intentionally don't pay their suppliers because they know their suppliers are going to go bankrupt and won't actually be able to see them or Amazon will be able to then buy that a supplier and bankruptcy for even cheaper. It's pretty predatory. So there's an alternative to this is where you effectively borrow against your invoice. So right, so your invoice right has some kind of credit rating. So Walmart and Amazon are, you know, very good credit. So there are good for paying out their invoices, and so banks are willing to extend loans using these invoices. Is Collateral. The only problem is that the banks won't deal with you unless you're doing like ten million, twenty million, fifty million. Sucks, because you're just not worth it to them. So there's a whole the most and most supply like manufacturers and suppliers are much more smaller than that. So what we think we can do is we can have trade trade shift are is the one that already does the trade finance stuff. So they already know what the credit rating is of all of them and you know risk is of all these different invoices. But they can do is they can tokenize this stuff, they can put it in a CEP, they can draw a die against it and they can convert that die in a dollars and they can wire those dollars to that supply company. And what you've effectively done is the supply company doesn't need to know anything about blockchain, they don't need to know anything about tokenization or die or anything. They don't even know they're using it. They just know that when they go to trade shift, if they like, sign over their invoice, they get dollars wired to their account within twenty four hours. Okay, that is power going to get amazing, and I'll tell you why. I come from a difference perspective. One of the main problems when it comes to anything dealing with independent contracting is exactly what you mentioned, just in a different context. So it's such is about suppliers. It's about, and anybody dealing with any organization which has a sixty day or ninety day or even six month sort of cycle for paying out your particular contract, type of contract, you are dependent on that cash flow. You genuinely are. Like you, if you want to be a small contractor today, you have to have at least three months worth of I can live off of this in your pocket right away. Otherwise you are putting yourself at extreme risk because the likelihood that you're going to get paid within ninety days is slim to none. Your probably even going to paid till after that because you have to actually do the work first, before the invoice is even considered valid, before you leaving get paid. Okay, so, like the barrier to entry into doing just independent work and independent labor and independent consulting, an independent anything, is tremendous, especially for young people, people who come from different economic situations, people coming into the country that is not from a privilege of a country. So people from India to here would have a more difficult time. A lot of these like problems are really easily solved if we just had a way of learning against invoices. But as you set they won't even look at your invoice for lending purposes as collateral unless it is at least probably one to one point five million minimum, depending on the back the job. Small, small trade, trade tree. Yet so I've gotten called excited. He's gone. If you can, if you can, not only, not only, if you could take these invoices,...

...if you can attach a credit rating to the invoices in and if you could somehow bundle the invoices together, providing an a sense of model for collective bargaining, you have something pretty big there. There that they have something pretty big there. That is a big, big, amazing feature that I've always wanted and it's only possible because of maker. I'm very excited. Well, I will just sorry, I go on rance to volive. That all as a ranter. It's okay, we love it. That's it's so scary, though, because, like, I guess it's not because it's not as scary becaus as we rebuild this financial infrastructure, we keep building on top of it, keep building on top of it, keep building on top of it. I've tracked away to the point where the users even know they're doing things. It's important to realize that the foundation is much more stable than reliable, or at least dependant upon a rational market, then something that is not or built on a house of cards. And so I worried a lot that what we're doing in the blockchain industry is rebuilding the financial like a more efficient financial system, but also the fact that like it may be more obscure, complex and and because we're building so many layers on top fit, that complexity grows and then it just the same damn thing, but also worse. Have you have you like thought about this type of stuff and how the future grows and we rebuild the financial industry? What's keeping us from not doing the same damn thing, and why is it going to be better this time? I'm worried about it all the time. Right, you know what, hate to beat the dead horse, but you know, two thousand and eight was was right. It was the result of dressing up a bunch of crap. And I see when what Collin just said. I if they use to build bold the market, I rebuild building these things up, baking them and selling them. If that's not done correctly, then we just do the same damn thing. So I think you made an interesting point earlier where, you know, we kind of determined that like make her super, super transparent. Right. So if we, if if we build on top of maker, we just have to make sure that the things were building on top of are just as transparent. And then we can we can actually say, okay, you know, this comes from, this, is connected to this in this way. The sooner you if you build something that's, you know, way too abstract on top of maker and that's not transparent in and of itself, that's where you start to collect risk. All right. So the whole point is that you collect risk as soon as something is to abstract. And so we all we can do is ensure that the bass layer is transparent. What we we can't control what anyone builds on top of it, right. And I mean I would hope that, you know, regulators would recognize risk or, you know, even better, that users would would recognize risk. But I'm I worry a lot of the time that's some APP is going to come out that's not isn't, you know, calculating the risk right. That isn't very transparent. And you know, they have some viral of growth and their users love them and it's going to end battle connect baby big gone next to exactly big connect likes. Yeah, and that's that's. That, to me, is actually an illustration of why these things are the amount of power we're giving people, like we're removing the the the curtain and showing that the wizard is not this big floating head and that these systems are something you can do and you can't. But the problem is people don't have the wisdom to operate them and when they're so easily accessed, you can get into situations like, like, like pozsie schemes, you know something, Barman schemes. You know, I think the biggest lie that the bankers were able to sell like the world's population on was, hey, this stuff is way too complicated for you understand, let us take care of it. Yeah, but why I worried about is we're building systems that actually are right, try explaining pitcoin through the average, everyday person and then then think about how more complex maker is as opposed to pick one. And then that's just the base layer, right. That's that. That's I mean, that's that's the technology side. So when what he's saying, that what I hear when he says, when Nick says, that is that Managing Your own finances is too complex. Let's do that for you. So so what? One thing, though, is people don't need to understand the technology, but they do need to understand the principles behind it, the risk, and understand the risk well. So, for example, maker, yeah, there are, you know, some risks that are specific to the technology stack, but most of it is do you trust the economic model, and the economic model we have an maker is not something new one. It's actually...

...something very old. Central banks use monetary policy, right, you know, modifying interests, race and and and the like as a way to you know, stabilizing and curb and platient all the time. This is really just the blockchain representation of that. This is not something new, right. When a you know, hedge funds use basically insurance kind of derivative products all the time to secure their positions. Right, the MKR, like as a insurance mechanism, is not something new. Right. What we've just done is adapted kind of existing models that work and just put them on chain in a more transparentment. So that's great and all, and you say it's the same, but at the same time, imagine getting these financial analysts who are up there as talking heads on Bloomberg, we're up there at CNBC. Who Up there on Fox business talking about how the markets moving and how things imagine them talking about maker do out and how make it do out, is positioning self and whether or not it is going to raise the interest rates are lower. And they're talking about it like it's some monolithic thing because that's all they know and they don't understand that there's a there's a mechanism in there that that that's out of the control of any sort of decisionmaking process and they you and some of them, will understand it. They might use the microphone to manipulate the markets so one way or another, swaying the mass opinion in the tragedy of the Commons. But either way the story is still changed because it's not the same as like the fet is going to change the rate. You know the it's the bet. It's a matter of it's a matter of who has control the power and it's just going to be an interesting world. When I see the maker market being up there, spoken right next to yas some peak, I'm very much looking forward, or I will bet you are. All right, we gotta I got a few minutes left or else sucks about maker? What else sucks about maker? Oh Man Um. I would say another risk we have is the mkre token is not super liquid right now. So I think that's something that we need to work on. So, right, so when new MK are gets diluted to cover a bad debt, right, that M kr needs to be sold. Right. That's how you that's how you can get if the markets for buying them care are very liquid, then you can see the price to apreciate rapidly rather than slowly, so the market cap of so you kind of see this in crypto right, where you have coins that are worth like, say, like a billion dollars, but they only do like twenty million dollars of actual volume. Preday, after you filter outwash training, it's probably like a ten of that. If you were to dump, you know, five million dollars of you know coin onto something that only has a million of real liquidity, right, you're going to depreciate the price extremely fast. You'll see that one billion dollar market CAP, you know, token, collapse to eight hundred million or, you know, seven hundred fifty billion quite quickly, and that's something that we need to improve about the makeup cooken if it's going to be this robust insurance product. And I think the solution to that is really to just publish more types of academic papers detailing how the maker system works, because modeling the value that the maker token should have is I'm not I'm not going to say it's easy, but it's pretty it's pretty intuitive. Right, so every die that's outstanding as some kind of fee that it's returning based off whatever collateral it was created with. So right, so that's that's money coming in. Okay. So you know, companies are valued based off like profit to earnings, right, you know how much, how much money they're bringing it? Well, you know, I'm K are is bringing in this amount of cash this year. And you know the die supplies growing, you know, seemingly at some rate x or on some kind of function X. and so we can predict what, you know, the castles will be three years from now, five years from now. You know, we can predict like, Oh, well, we're taking this much risk in terms of something bad happening and US having bad debt and needing to have some dilution. Okay, so when you have some factor, for mkr has this probability of getting diluted. But what you find is that you quickly, very quickly arrive at, you know, a numb out, a number of what Mkr should be valued done. And as soon as you have that kind of clarity in the market of what an Mkr should be worth, given the state of the system and given kind of the where the system is trending on going, there should be a lot more buyers and sellers who are willing to sell slightly...

...about that price and by slightly below that price. Yeah, I agree. I think that the lack of modeling and data for people to look at definitely keeps that velocity down right. People using the token and stuff like that. Yeah, we're working on that. Other people are working on that, and that's just that's also a growing field which is dependent upon the actual system being in place and work right. Can't built bottles of that data a lot of its, yeah, or you can't validated models, but data. Well, rather than close on you know, something that a question of like yeah, what needs to be fixed? What is what is your favorite thing that you've seen happen in the maker ecosystem that just blew your mind or would just like was you know, emergence that you didn't expect or that you are just super excited that you actually saw realize what was what was your favorite moment? My most favorite moment was when we launched die December of two thousand and seventeen, and it actually worked. Only seventeen. You know, eth is that like eight hundred dollars. A month later at rallies to like, you know, on fourteen hundred dollars, and then you have a year of just crashing, crashing, crashing, crashing bear market. It was really the worst possible time you could launch a stable to it to can like die that, you know, was dependent on ether to back its value right, and ether with blocking, dropping, dropping, dropping, dropping, and in that entire time and even up to today, we have never, ever had an underwater loan, not once. And that is that is because in our parameters, you know, on how we you know, that minimum collateralization ratio of eth like. Yeah, sure, each heath drops twenty percent some days, but it doesn't draw up fifty percent in an hour or it's not very likely to drop fifty percent in one hour. And so that's what keeps the system solvent and that's kind of what gives me so much confidence. That's all a testing it to the to the dynamics of the system. Right like that, how you can help? How quickly you're able to move back to equilibrium depended up on the price changes of various things. On fourteen hundred dollars down to eighty. It's pretty solsible. Like. I mean, would that only improve with you add more assets system? So if you're backing Bitcoin, is that like basically heading your bet off of the theoreum? Is that going to improve the rebustus of the system? Or do you see the of course, the more assets you bring in, the more stable, the stable clim becomes. Yes, I'm going to challenge that. One account of if I flash crash is okay, you at least have a reference price over here. On other another o the coincident that that's something, something's up with the theorium and it's not too big of a deal. You can then suddenly implement measures. Like the flash crash happened only on ATHERIUM, and that was a pretty impressive flash cut. I mean dropped down to what eight cents from like, who was it at the time? Like three hundred, and that didn't affect us and you want to know why you were open for that? That's right. Now, we we pull our like Oracle prices from a bunch of different exchanges, and so each or each oracle itself, right, is pulling from, say, like five exchanges and it takes the volume weighted average price. said, it knows to filter out outliers. When it's just one exchange that just goes tits up. So we actually weren't affected by the flash crash at Ale holy. So, yeah, I don't know. So why would you say you would argue against it, cory, because I've had a lot of advanced courses and reaction dynamics and this is probably something that models are varies very well to that and it's a subtle science and there's could be compat some kind of like it. It's a as you increase the collateral, it becomes a larger and larger multivariate problem. As with any multivariate problem you have exponential increase of complexity as you increase the number of variables and optimization of those types of problems gets incredibly complex and sometimes non trivial. So like like in terms of like the dynamics and what can what can pull that system out of equilibrium. So I think that be that be kind of like a area of research that I would love to see, because the majority of modeling we do in token economics is usually ways for firms to do to make appropriate decisions on like collateral, like where the we're they're putting their money, basically making good quality buys decisions, things like that. But modeling these things, especially when things have multiple utilities, with more complex theories that may be worth wall, like ecology theory or reaction dynamics theory, may be more appropriate it. I'd be kind of curious to see like how those play out. I'm interested too, I think. I think some of the things you brought up actually bring to mind other questions, like are the fundamental assumptions werere making on how supply and demand actually works on a system...

...like this sound in the long term? Im enjoy the time they use like that old, old model, like in with a pqekuse Mv Right, and then what with? The major of the problems are modeling velocity, and so you have these like really odd ways that don't quite fit on modeling these types of like old financial models that then maybe make decent approximations, but it may be out of the context, because the financial model just doesn't quite capture the complexity of the token itself right because at the end, like cryptocurrencies are more general, they transcend the idea of old financial old financial tools, and when you use the tooling of the older transcended thing, they're never going to capture everything and that complexity may mean dress the consequences of like the thing in practice and until we come up with better models or try different things, than I'm not I'm not too sure of like how accurate these things are. But that's that's part of the field and it's a very, very small, budding field. You need a lot of work and a lot of research to apply of these things and it's just a small amount of people doing it. And you know, I feel like it's also one of those things that sometimes you won't know until the bad shit happens. But I mean so far you sustained through all beetle, brutal beating, beetle brooding. You can give you a sustained through a beetle brooding, a brutal beating. And you know, it's been a rough year for markets and they're now on the upswing and you're watching the other side of it. So you know, I mean that gives nothing but confidence to what you're building. I just that's all I got to say. Is Your contracts held. You know, your contracts held, your system held through a rough time. I'm sure people are trying to come at you all the time. I guarant I guarantee there's people who want to figure out how to break you just for fun, not even because they have an edge, because that's what you get. Two percent of either locked up the definitely not doing it for fun. Well, I might be doing it for fun, maybe like like lock, like let's say that. You could just want to find a way to just lock up that money like that would just be because their dicks. It's done. It's like, you know, like I don't see in your maybe not. Maybe they want to tack you through itself, but I don't know. The point is that people are coming at you, you're you're standing up to the bullets and it's panning out so far and you building quite a hell of an ecosystem around this and I just very impressed with what I'm hearing here. So yeah, I came in as a pure skeptic intentionally and I came out as a person who is venty five percent there, because I always believe that that little, that little bit extra, because I got more to learn and, just like so much, you can learn it to ours and talk it. So but yeah, I know, I think this is really impressive. Stuff like this is extremely impressive. So cool. Well, I'm glad. I'm glad we got you seventy five percent of the way there. I gotta leave that room, man. All Right, let's go ninety. I'll go the full undred and ninety whatever, but like I got also kind of like leave myself, like you gotta leave yourself space. Like this is not an endorsement, it's just me trying to figure out what's going on here based on what you said and what I've been reading and what I've been looking at. It's it seems cool and reasonable and like it makes sense when you explain it to me the way you explain it to me, because before coming into this it didn't quite and it's just like there's a lot of stuff going on here and it's like I can't crock all this and this makes a lot mix it a lot better. The problem is that there's no clear entry point where you starts and then you go be to see the D it's all the thing and it's connected to this thing and this other thing, but we can't talk about the other thing. Let's talk about this first, and we just try. We just tried to do it over the past two hours. You have the whole picture that you see how things are interconnected, and that's why Corey was like, Dude, if you if you haven't listened to the previous episode, you got to listen to it before you do this, because I feel like you know just that, that that explanation of the it just helped. So yeah, I mean this is your point of Entry House. That push us really hard. I just tell everyone they should listen to these two episodes. It's before they even talk to Y'all, and we'll make it, make it a deal. In terms of like getting call of the rest of the ten percent. Where do people go to find out more and to learn the things that they maybe quite didn't get the first time when they listen to this or they need more clarity on certain things, for where's the conversation where they learn more? Where do they find you? So I mean if you go to chat dot maker nowcom we are very, very active on our chat. We have a super superactive community. You know, all of us are posting in the public channels like all day every day, so you can ask all the questions there. We also have a rebo on github in the maker down Rebo called awesome maker down, and it's pretty awesome.

It is very awesome. It is a list of like every single thing having to do with maker, with how it works, with kind of guides, with specs, with FA ques. It has links to every AMA, every podcast you've ever done. It has videos, it has links to every platform that trades die, that uses die. It has a ton of different stuff and it's just growing day by day, by day by day. Awesome. That's a lot. We really appreciate you take an extra hour and up to expand this out, because it's just there, just so much here and I feel a lot better about my understanding of how things work and I also feel about a lot about better like pointy people in the right direction. Would they have questions for me? Thanks Lot, thanks awesome. Hey, thanks for having me, guys, this was this was super cool.

In-Stream Audio Search

NEW

Search across all episodes within this podcast

Episodes (118)