YouSaid · the spoken record

Charles Hoskinson

lines on the record: 392
first: 2021-06-16
most recent: 2021-06-16
sittings or episodes: 1
sources: podcast

Every line below is reproduced as it was said and linked to the record it came from. Nothing here is summarised or generated. Directory · Search · Corrections

“It's how you build your career, too. There's plenty of people that they've gotten tenure just kicking the hell out of Intel SGX. Go to CCS every year. There's some guy there, and he's having a hell of a time making Intel cry.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Oh, that's pretty cool. So that's the other point about doing things in a very rigorous way that way creates a lingua franca for what you're trying to solve with the totality of the academic community. So suddenly people that you've never met, you know nothing about have read your papers, cited your papers, and start writing their own papers, either to try to attack and destroy things you've done or to build on top of the things that you've done.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Exactly. And, you know, and we've now gotten to a point where we're starting to work on refinements to the system rather than fundamental things that are like, if you don't solve it, the system just simply doesn't work. For example, you can run all this with NTP as your clock server, but you actually can create a notion of time within. We wrote a paper called Warborse Kronos for that. That's not necessary for the system. It's just a nice to have thing. It's a nice property. Optimization of the random number generation is another example of that. You can run it with a heavier thing. You just have more blockchain bloat and slower time and transition. We have this concept of epic. So you elect leaders to run the system every five days with Cardano. But there's been derivative work. We didn't even do this. This work occurred at University of Illinois. And that derivative work said, well, you don't actually need to do that. You can do it on a block-by-block basis.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“You put all together and you have this Frankenstein paper monster, and we're like, Where do we submit this? Where does this go?”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“No, it's the starting. And then I was getting back to my original point that you build things in iterations. Every step, if you've done it right, is an invitation for 10 more sexy, fascinating, fun problems. And this is why we have such a great time building labs. You know, we started in Edinburgh. Now we're at Tokyo Tech and University of Wyoming and Athens, and we're setting up more labs this year. And all these academics want to work with us. Hey, because we write a lot of really fascinating papers, but B, because we're focused on all these really cool, sexy interdisciplinary problems. We're actually running the problems where we don't even know where to publish the paper because you'll have this paper where there's like these PL guys working with crypto guys, working with systems guys, working with economists.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Yeah. And so that's a complex thing, and you have to actually model the game theory out to understand where those parameters should be set. And we didn't know how to do that. So what we did is we bought talent. We went to Oxford and we hired this guy named Alias Casupis, who's an algorithmic game theorist. We said, hey, would you like to do some game theory work in crypto? And he's like, that sounds fun. So he spent a year and a half. We built all these beautiful models and we kind of figured out what those curves needed to look like.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“So it lived there. And then the stake pools weight in the system is proportional to the amount of stake that they have delegated to them. And then we have this other limiting factor, K, which says that you get diminishing returns with the more stake you have. So it's kind of like an S function. So you kind of go up and up, and then eventually caps, and then at some point you get no rewards beyond a certain threshold. So there's an incentive to split pools to different owners after some points.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“It's kind of fascinating. It's a super simple concept. So you register a pool and then the pool is there and basically they advertise and they're actually registered on chain with a certificate. And then in the wallet software itself, you can see all of the pools that have registered. There's over 3,000 of them now inside the system. And you can click a little tile and it shows you all the metadata that's in their certificate and says, hey, I have my own pool. It's called rats, king of the rats. So you can see all this stuff that's described there. And pools have an operating fee because they're like a business. And they say, well, if you delegate to me, I'll charge this much. So if you get like 100 bucks in rewards, I'll give you 90 and I'll take 10 or something like that. And then you make your decision. And whichever one you select, you click delegate, push the button, and then you have now given your staking rights to them until revoked.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“System, but you're only as good as your weakest link. You're only as capable as whatever the basic cell can do to a specialized system where you start having these actors in the system that are actually a little different than the other actors. So you introduce this concept of the stake pool. And suddenly now you have this actor where you're probably going to be online 24-7. You're probably going to have extra relay infrastructure. There's a trust relationship where you don't own the ADA, but you have a right to use it for something. And a person's made that choice to endow you with that. The minute that you introduce specialization, though, the system gets more complicated. The game theory gets more complicated. And then you start having to think really deeply and carefully about, okay, well, can this now introduce a new attack vector that we didn't have before?”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“A system that would actually work and is working, that's what's implementing Cardano today, $50 billion cryptocurrency and all these people. That was a huge leap. But that paper alone wasn't enough. We also had to layer on the economic model because we said, well, hang on a second here. Not everybody's going to be online all the time to be available to make a block. So you need some notion of delegation. The minute you have a notion of delegation, you have these stake pools. What the hell does that mean? And so this is a beautiful kind of interdisciplinary notion that layers computer science and biology together. And minute that complexity starts going up, you start seeing cell specialization. So you go from single-cell organisms to organisms where you have eyeballs and brains and hearts. And each of these tissues do different things. Well, analogously, complex distributed systems start getting specialization. You move from the single cell thing, Bitcoin, where everything's a full node. They all have the same rights and responsibilities. A lot of homogeneity in that.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Yeah, not only are they really hard problems, they actually require different cryptographers because you're moving from mathematician-style cryptographers like the Neil Koblitz and the Addis Shamirs and the other people that start as proper mathematicians. They really love theory. And that's their thing. And the proofs are dense and they're thick and they're beautiful to practical applied work where you're saying, okay, now this is something an engineer can look at and say, I know how to build that. I know how to think about that. So that transition from GKL to Ouro's Classic to Prause, I'd say the biggest leap was Classic to Prase because that was going from a system that would only work in a consortium chain like fabric.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“So adaptive security also, we move from an MPC random neighbor generation, which was great, but very heavy and very slow, and you can't scale to large amounts of people to a VRF-based system, which is super fast, but a little dirtier because Al Gran actually did some great work there. There was some good knowledge there.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“What actual network conditions are you looking at? Are you okay with living with an external clock or do you want to build time from within? How are you generating random numbers, et cetera, et cetera? And every step of the way, each paper you're solving one particular class of problems. With PRISM, it said probably shouldn't know ahead of time who Eve is. You probably shouldn't know who's making those blocks. That should be something after the fact. But if you know ahead of time, you can attack them. You can DDOS them. You can cause all kinds of problems.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“The papers that we published, and they're all peer reviewed, like GKAL was Eurocrypt. That's a very hard conference to get into, and Orbor's Classic was crypto, and Praus was EuroCrypt, and Genesis was CCS. So basically every step of the way was first an academic validation that there was some merit to the work that was done. Second, it solved a particular class of problems, either showing the feasibility of the entire problem. Because when I said, let's do the model first, because let's see if we can do an FLP thing. Let's see if we can get a possibility theorem. That's great because you're done. It's like those short math papers were like, I found a counter example. It's like, oh, okay, this whole thing has fallen apart because you have a two-line proof. Thank you. So that's what we were looking for in the beginning of the agenda was let's either prove it's possible in a straw man case or show that there exists an possibility result in which case we can just abandon the entire inquiry proof of stake's impossible and then once you've gotten past that threshold it goes from theory to practice”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“What is this thing, right? What's the security properties this stuff? Like, what do we really mean? Then we did Auroboros Classic, the original Auroboros Protocol in 2017. And that protocol was like a synchronous system and it assumed the nodes were always on. It worked, but it was useless because that's not real life. Then Prauss came out and then suddenly we relaxed things”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Yeah, this was a philosophical difference between Vitalik and myself. You know, the problem with the people in the Ethereum side is they're really bright and these really bright people, what they do is they try to do everything all at once because they're really, really smart and they keep going until they run up against the wall and they realize the problem is a lot harder. If you're more experienced, and that's why we brought in proper academics like Agolis and others, because they'd been beaten up through life, you know, Agalos worked with David Chom and these other, it's really hard work with those guys. And they'd already been humiliated and yelled at and had chalk thrown at them and all that stuff. And so they were humble enough to say, I'm not smart enough to solve the big problem. So don't even try. What you do is you decompose it and you say, okay, what's the first problem to solve in a chain of problems that you can compose your way up to a working system? And once you get far enough along, you have something that's pretty good and then you have an obvious path forward of how do you iterate and improve that.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Do it. So then your system will centralize along the weakest link, whether it be the storage layer, the computation layer, or the network layer of the system. So if you can incentivize the resources differently, then you'll be in a beautiful position where you end up having a resilient system that pays its own bills.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“And they can make just as much profit mining this as they can make mining the other system. And the markets allow them to profit from the destruction of a system. So that's something that proof of stake doesn't suffer from because the only way you can participate in a proof of stake system is you have to actually own equity and you have to have ownership in that system. So if you go and destroy Daisy's chain, it would just be a net loss for the most part, unless you have really messed up markets or something like that. So there's always trade-offs in all these things. And this is why I like this concept of going one to end and having multiple resources, because why not have proof of work and proof of stake together if the proof of work is useful, not wasted computation, and why not add other things like create incentives for network relay? Right now there's no incentives in the system for you to run peer-to-peer nodes and share data. Right now, it's not a problem, but if you're running like Amazon web services level of bandwidth, it could cost you like $5,000 a month in bandwidth just to run a full node or something like that. No one would.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“And they use the same algorithm. Then the problem is if the miner comes in, and let's say the miner has enough resources to have 51% for any of these chains, they actually have a perverse incentive to come and destroy one chain and short sell the asset. It's called a goldfinger attack, and then go mine the other asset because they're not bound to that asset. They're not loyal to it.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Talked about better network conditions. They wrote a paper called Sleepy. We did Genesis. We also did the very first Pruly Secure Protocol. But that was six years of work and like 12 papers. And it's still not done. There's still a few polishing things that have to be cleaned up because this is a physical resource. And there's something there. But there's a flaw to proof of work that is a little problematic. It's a winner-take-all type of a system. Maximalism is kind of philosophically and computationally built into it. Let's say you have And they have roughly the same market cap and hash rate.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Network and Wally joins the network, and Daisy joined the network, then you give them some different chains, like five or ten different chains, they can run a calculation and they will always pick the longest chain, the heaviest chain inside the system. That's a great property of proof of work until we publish Ouroboros Genesis in 2018. You actually needed to solve that in proof of stake with a trusted checkpoint. So some actor had to be observing, watching the whole thing, and creating checkpoints. And then when new people joined in, they would only be able to distinguish between a chain based upon a checkpoint telling them that. So you have to do a lot of really wonky, crazy math to show Crit's notion of density to be able to show that that's possible. But there's a lot of properties of proof of work that were super hard to replicate and emulate in the proof of stake world. Macaulay kind of revolutionized the whole VRF thing. There was a group out of Cornell that”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Yeah, it's a great question. And the first question we had was, could proof of stake actually work or not? So the problem was that the engineers kind of led when the science should have led. And so there were all these POS protocols that came out in the early 2010s, like Purecoin was the first, and then NXT and others came out. And there they had suffered from things like the random number generation wasn't good. They had grinding attacks and nothing at stake and all these other things. And there's a lot of beautiful properties for proof of work from a theoretical sense. We even wrote a paper called GKL named after the authors Juan Garay, Nico Leonardis, and Egolos Gassis, our chief scientist. It's got 1,100 citations now. It was published in 2015. But basically all it did is just modeled a blockchain and created some security properties for it. And then it started talking about, well, what is proof of work actually do for you? And it turns out it does a lot. It's an asynchronous system. You can bootstrap from Genesis. So if Eve joins...”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“WeChat just like trying to sell miners or trying to figure out how the hell do I move miners, you know, because they have these huge data centers they've constructed. You can't exactly go and grab a server and like take it with you. It's huge. It's a lot of work. And if the government seizes it, well, it's their property now. A virtual resource, you can click a button and redeploy it to a different jurisdiction. So to me, for a virtual asset, it makes a lot more sense to try to tie your security to something endogenous, something within the system, because it's just like the asset. It can move anywhere at a click of a button. And human beings have a much harder time attacking something like that”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Of the shares. Now he has less than 5% of the shares. So there's founder drift over time as the value goes up, divestment occurs. You have more and more and more people coming in. That means there's more people who can participate in the consensus. You can even tune economic parameters. And this is what we did with Cardano and Oroboros. We created this concept of K in the system and it's just a parameter. And it's like a forcing factor that tends to accumulate a certain amount of stake pool. So you can set it to 200 and then 500 and 1,000 and so forth. But the basic idea is the price of eta goes up. You make k larger and then you end up in practical terms having a larger and larger set of actors making blocks that are unique and distinct. And the other good thing is this is a virtual resource instead of a physical resource, which means it's portable by a click of a button. So let's say China says mining is bad, we're going to shut it all down. It looks like they're moving in that direction. You have all these people.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“The system becomes less decentralized over time because you have these vertically integrated operations. I mean, not everybody can go build a mining facility on a volcano in El Salvador. Not everybody can go to Mongolia and set up a five gigawatt power plant in a huge data thing. Not everybody has access to the patented ASICs that people produce. Because what if I don't sell it to you and I have the patent on it? Or what if I control the supply chain for these things? So you'll end up having centralization around maybe 10 or 5 major operations, as we've seen historically with proof of work. And that means you end up having a ruling class of a mining oligarchy in the system. Proof of stake, if you design the parameters correctly, you actually get more decentralized over time. Because as the currency goes up in value, the distribution of the currency tends to get more egalitarian. For example, Bill Gates, when he started Microsoft, he had 44%.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Afihar's work and Yonatan Samlopinski. But the basic idea is you pick some collection of people, they make some collection of things, and there's some way to sort it all out, serialize it, and prevent double spends. Great. Proof of stake is the same, but it's a synthetic resource. So instead of doing things, they say, well, if you had 25% of the hash power on average over a long period of time, you'd probably win 25% of the time. Well, why don't we just introduce some randomness in some source and then 25% of the time on average over a long period of time, you'll win. So it's a synthetic resource. But you still have to do the other two things. You still have to make the block and you still have to validate the block. The big difference is this step in the proof of work world is horrendously expensive. You use more energy than the nation of Switzerland. And the problem with that is that you have less resources for the other two. And the other problem with that is that if this is horrendously expensive, you have an economy of scale kick in. So what ends up happening is...”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Okay, so they all have the same three properties pick someone in charge, do something, and validate it. The difference is that the picking mechanism for proof of work is you have to solve a puzzle. So it's basically like buying lottery tickets and you can buy a certain amount every second with your computing devices. And some of them are ASIC resistant, so you run them on like a laptop or a GPU, and some of them specialized hardware that you have to either manufacture or buy from someone who sells it to you. And that's just how many tickets per second you can get. And eventually you hit those magic numbers. And when you do, it means you have the right to make the block. And generally you bundle the block making with the proof of work system. Now you can do this looking for a single or you can do this to actually shard it and look for multiple block makers at the same time. So there are sharded proof of work protocols like PRISM is an example of that. And actually Ethereum got started this way with Spectre and Ghost and Phantom.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Any of these other guys, but now the next step is say, well, why don't we just have one? We should have multiple resources. And the point is each of these has different trade-off profiles. And so they balance each other and you end up building a much more resilient system. So it's not win or take all with one particular demand.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“But basically, you have to figure out okay, well, somebody has to be in charge, some group has to be in charge, and you can do it with a meritocratic, hashocratic computation thing. You can say, well, if you have coins, 25% of supply, 25% of the time on average, you'll be selected to have the right to do this or give it to somebody else. Or you could search for other resources. And they can even be human resources, like some notion of merit or social benefit. Maybe you get a token for that and you can weight it with these other systems. And that's kind of where everything's going. You know, we're getting to a point where we've really optimized all the properties here. We've proven all these nice things about it. And there's a lot of competition to basically build like.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“So that's really the crux of it. You have a ledger, and the ledger is just all about saying, hey, we need to put some stuff in here. And once it's put in here, you can't turn it back. And you know when it was put in and everybody can see it or some group can see it. And then you need to pick somebody to modify that. So all this chaos will happen, all these transactions are all around the world. Our perception of them are different. There's a beautiful paper from Lamport that kind of talks about this from the 70s. It's like one of the most classic papers ever in computer science, I think. It's been cited like 50,000 times or something like that. It's crazy paper.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Hybrid proof of work proof of stake. So some of the blocks were made with the token ownership distribution and some of the blocks were made was proof of work. But you could keep adding. You could put in like, hey, I want hard disk in my thinking. You put permacorn in or something like that. So create incentive for hard drives. And then you could say, oh, no, I want to do like a human system, like a proof of merit. Oh my God, now we're up to four. And you just keep adding. And each of these pools will have different adherents and actors. And then you can actually balance out the whole thing.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“So there's checks and balances in guards all the way through. There's a meta question of fairness in all of this. So the proof of work people, they're kind of a cult. And they say that this is the only truth and everything out here, any other resource is not legitimate or valid. And there's not a lot of evidence to that, but that's what they believe. The proof of stake people, the downside and weakness they have is it's a plutocratic model. The more ownership of the system you have, the more control you have over that system. And it suffers from the same thing the shareholder models suffer from, whereas you may maximize short-term gain over the long-term viability of the system. So a really cool question is, can you build systems that are multi-resource? So instead of just pulling from one resource to select who wins, this 25% of the time and maybe this 25%, you can do that. In fact, the cryptocurrency space did that a long time ago. There was a cryptocurrency called Purecoin in 2011, and it was a high-level.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“There are other ways to potentially conceive of this, but this particular model gives you a kind of a way of thinking of all consensus algorithms in one setting. You can be algorond, you can be a classic BFT protocol, you can be Paxos, you can be RAF, you can be proof of work, you can be proof of stake. It's always the same idea. You have to find someone or some group to be in charge. They'll reach a consensus on order. They have to then do some work. change the state of the system, update it, and then the network has to accept that that's valid. So even if this process works well, this side will say, oh, you created a Bitcoin out of thin era, you're not allowed to do that. So that's rejected.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Right. And so, anyway, the first step is all about basically deciding who's in charge for that moment. So blockchain is just a sequence of events. The heart has to beat, the metronome has to click. So somebody has to be in charge. And so generally you have this notion of a resource. So there's some pool of resource out there. And it can be a token. And in that case, it's a plutocratic system. And that's what proof of stake does, or it can be computation. But there can be other resources, but computation is what proof of work does. And so you make so many hashes and then eventually somebody wins. And that person who wins is now the person who basically gets to decide the order of transactions and put them all together from their perspective in the system. Then once that person wins, they'll make the block. That's step two. And after it's made, transmitted and it gets validated and accepted. So actually it's quite fortuitous. You have the magnifying glass because at this stage people are trying to decide is what I'm looking at correct or not.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“That's the whole point of a consensus algorithm. So, whatever ledger you're running, there has to be some mechanism to decide who's in charge. And that's what proof of work does and proof of stake does and all these other systems. And you break them down basically three steps. And so we'll use Eve for kind of step number one. Hi, Eve, how you doing? And we're going to use...”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“But they have relatively the same mechanics in terms of those properties of auditability and timestamping and immutability. You know things won't be reversed, you know, that people aren't going to manipulate the timestamps and you can audit at least enough to know that the ownership is right.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“But I needed to treat him because he's quite ill. They'd say, Who is this person in Zimbabwe? I don't know you. I can't give you his records. I need his consent. Oh, no, he's unconscious in the hospital. Can't do it. Well, a broker system that would allow the movement of medical records would be an example of what a blockchain could potentially do in the foreseeable future. Cryptocurrency is just an application that runs on top of blockchain because it turns out that when you issue property, you also can issue tokens of value. And then you could have a monetary policy. It can be inflationary or deflationary. You can have demarage where it decays over time or whatever have you. And the very same mechanics that would ensure your property records are secure, your medical record access is secure could also be applied for the ownership of the cryptocurrency. And again, you can either be completely transparent and everybody can see what everybody owns. And that's what Bitcoin does, or you can be as opaque as you seek to be. That's what Zcash basically attempts to do. It says, hey, let's keep these things as private as possible.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Okay, so that's the first part of it. The second part of it is that it's Auditable, meaning that instead of saying only the high cleric or the president or some very special club of people get to see what's going on, suddenly now all the people can actually see who owns what, where. Like imagine you had a tax system where the pro-propocate just leaked the taxes of all these different billionaires and said, well, how much do they make and how much they pay? Imagine a tax system where that's just done by default or other social systems where this type of information is put in by default. So it's tremendously useful, this type of structure and all kinds of things. Medical records, supply chains, just a good thought experiment is I travel a lot into 52 countries in the last five years. Imagine if I got sick in Zimbabwe or something and I'm unconscious and in Zimbabwean doctor calls my doctor in Colorado and says, hey, I need all Charles' medical records. He's unconscious right now.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“The political manipulation, and they're brittle to geopolitical events. For example, when Syria fell apart, you know, the very first thing ISIS did is they started saying, hey, the ownership of the land, it's going to fundamentally change. We've decided that this guy over here now owns all these things. And then when peace comes, like, how do you unwind all of that, put it all back together? So the power of a blockchain is that it gives you a transnational way of sorting all these details out, putting all together in a place that you know that even if it's inconvenient to a very powerful actor, that it will still stay preserved. This is an asymmetry we haven't had as a society, usually kings and empires, they have the ability to decide what's true. And then suddenly you have this asymmetrical thing that is above them, kind of like a synthetic laws of physics. And once something goes in there, you know that that's there.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“So blockchain is just a ledger, and really, it has three nice properties your timestamp, your immutable and auditable, either in a global or a local sense. And so there's all kinds of things mankind has invented where it's really important that when you put some information down, it doesn't change and other people can see it and that you know when it was put down. For example, a property ledger. So when you buy land or you have rights associated with land like mineral rights or water rights or these things, you'd like to transitively see how does it go from Alice to Bob to Charlie to Jim and so forth and what was the state of these things as they were transitioning. So how much did they pay? When did it occur, et cetera, et cetera? The metadata that follows that. Okay, well, normally these types of ledgers are so important that they're managed either by governments or regulated entities. And the issues are that while they can be efficient, they're generally brittle.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Well, that's the beautiful thing. When you have a less expressive model on chain, it means you can do anything you want off-chain.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“But there you'd want to have the whole round trip. So you'd like to have an interaction. And I think a video game would probably show it the best. Like if we could re-implement crypto kitties or something like that on it, and you have this off-chain infrastructure and you have your GUI and your front end, it's running on your phone or a browser, and most of that lives off-chain. But your crypto kitties, they'd live on the blockchain. The whole round trip end-to-end with relatively low fees and low latency and high availability of service. It never goes down. That would probably be the best thing to do. And we'll have something like that by August. It's pretty easy to build this stuff.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“You want to start putting peer-to-peer dynamics inside your system, you're going to gracefully connect to that on-chain code. And it's very clear how those two things connect together. It just so happens Haskell is really good for this. They have template Haskell and it makes it very easy to embed domain specific languages and it makes it very easy to wire your Haskell code onto off-chain infrastructure. So in the future, you'll be able to have your off-chain run a node or the job of virtual machine or a.NET application and there'll just be this beautiful interface and then it can talk to all your on-chain code and that's written in that DSL and you have a high degree of assurance that it's right.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Goldilocks zone. Bitcoin script was too restrictive, and every single time Satoshi tried to dial it up, it led to mega problems. Like there was a beautiful thing called the value overflow incident in 2010, which led to the creation of billions of Bitcoin. They had to quickly clean that up and sweep it under the rug and pretend like it didn't exist. But that was mostly because of an issue with how the scripting language was implemented. When you look at Ethereum, it's like this pyrrhic game of stomping down these skirmishers where every update there's something they have to change or tune and then it's not clear how you shard such a model. So we said let's build something that's in the middle of this and that's what Plutus basically is. And it's really designed to play very nicely with off-chain infrastructure as much as on-chain infrastructure. So you can look at all those different examples, whether it's Wolfram wants to auction off their universes or Blizzard wants to issue an in-game currency or you're Uber.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“But then you've opened your application up to a broader world. And what we've done now is added a blockchain layer. And the blockchain handles the accounting of that asset and the spending policy of that stuff. So that is a much smaller program than what Blizzard is doing with World of Warcraft. So the point of Plutus was let's create a language where you can write these small to mid-sized programs and have a high degree of confidence that they behave with correctness and also they give you deterministic results on the consumption of resources. You can run things locally and you actually understand what it costs to run and that doesn't change when you deploy it on the system. If you dial up the expressiveness of the system and like Ethereum does and these big mutable account systems, the problem is you have to have global state. So whatever you test locally doesn't actually necessarily translate to what you've deployed. Okay, so we spent a long time asking where's the”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“So, Plutus, a programming language. It's kind of a DSL that we built on top of Haskell. And basically we wrote it after spending about three years thinking about smart contracts. We were trying to figure out what is the ideal language to express a smart contract. And then we started saying, well, what is a smart contract? Is it the whole application or is it just like a submodule within an application? And usually it's the latter more than the former. You can build a self-contained program like a script, but usually what's happening is you'll have it like a video game, let's say World of Warcraft or something like that. You say, hey, maybe I want to actually create gold in World of Warcraft that's actually a currency, you know? Okay, so I'm going to issue a token. Well, and then maybe I want to create some mechanics behind how people are going to trade that amongst each other. So that would be like a smart contract layer, an issue, an asset. So you have this centralized server running and proprietary software controlled by a single company.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“And so your main chain will always stay somewhat conservative and have the minimum viable amount of expressiveness required on it to do all kinds of interesting things and also for interoperability be able to talk to all kinds of interesting things. But it's not trying to be everything to everyone. There's never going to be an ice cream store in the island. You'll have the grapes and the beautiful women, but no ice cream.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“And it's not clear if that's the standard yet or if MySpace or BlackBerry or all these other things, it'll fade away. Well, if it becomes the standard, okay, don't fight nature. Just support it. And the same thing that gives you the ability to bolt on the LOVM will also give you the ability to bolt on the EVM. And they can run with their own models and they're encapsulated, bulkheaded, separated systems. But you can move Ada applications information between those two systems.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“on the island the ocean's big it has everything but the ocean's got sea monsters and sharks and bodie mcboat face and you know all kinds of crazy stuff Right, so that's what Yella is about. It's basically this, let's bring LLVM into our world, and at some point in the next three to five year time horizon, we can bring modern programming languages in, but they're going to come in with all their flaws and their warts and their problems. And then the pond was the idea of the Ethereum virtual machine. There's some networks.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“And that's why we work so closely with the University of Illinois about a champagne, Grigory Roshu's team at runtime verification. What they did is they said, let's start with something very familiar, LLVM, which has been around for a really long time, and they happen to have created it there with Apple. And let's take that and translate that into the blockchain space. Okay, then what you have it, then it's very easy to modify compilers of standard languages like the C's and C++ and other such things that do compile to LLVM already and have them run there. So that's a different execution model than what we tried to build for Plutus, which focuses on correctness. Okay, so then all you have to really do is say, can both of these models coexist within the same ecosystem? Because then you kind of, and I did a video, it was called like the island, the ocean, the pond. And the basic idea was say you have an island where everything's perfect. Calypso lives there. Life is great. People feed you grapes every day. But maybe you can't do everything.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source