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

“Yeah, so people want to create exchanges that don't have custodial risk. The point of exchange is to build you a marketplace where bids and ass can find each other. People can meet and trade. You can find a price and you can get liquidity. So you got gold. You want to turn your gold into dollars. Okay, well, somebody has to create a marketplace for that.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“But God, that's a huge coordination problem. And it's a huge information presentation problem and incentives problem and so forth. And unfortunately, people who value being first to market will kind of piss in the pool for everybody else.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“There's a bot that says, Hey, give me some money or something like that, that kind of stuff happens. But then just protocol level flaws, like what happened with the Dow Hack, that's what really keeps me up at night. How do you resolve that problem? Because I don't hire these people. I don't tell them what to do. I didn't tell them how to build things on the platform. They may have tons of experience and knowledge that could be Simon Payton-Jones, or they could have absolutely no knowledge whatsoever. And they read one tutorial and they've written three lines of code in their entire life, and they've deployed something horribly broken, copy-paste, and suddenly it all goes to hell. I'm judged by both. That's what really keeps me up at night. And we're as a company trying to figure out, and as the ecosystem trying to figure out standards, like a university, Wyoming, we're setting up the smart contract engineering institute. We're negotiating with them right now. And the goal there is just to create some standards for how to certify smart contracts so that you can get that green check mark and know that it actually has some assurance level behind.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Know, and that's the issue you get as a platform developer coupled with the wins and losses of your application developer. So if they go do amazing things, it's like, oh, yeah, Windows is great. We love it. And if they go do terrible things, they're like, oh man, I guess he's trying to kill all of us. What we're kept up a night about is not just what we've constructed, but also how do we curate an ecosystem and foster the development of an ecosystem where you have assurance baked into the application. And that's somehow expressible to the user. So when you download your smart contract or you click your one-click install, the user's Uniswap clone or whatever the hell it is that's deployed on Cardano, you have a little green check mark or something that indicates to you that somebody audited the code or followed a specification. The lack of that is problematic because then first there's impersonation, the whole myther wallet thing, or like your videos or my videos are the same. Every comment.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Exactly right. Messing your whole world up right now, Alex. No, I mean, there's so much that keeps me up at night. It's like you're in a cold sweat every day when you have an ecosystem like this because the other thing is you're judged as much by the applications people build on the platform as you are by the platform you've constructed. So one of the most unfair things that happened in our industry was blaming Vitalik for the Dow hack. He didn't write the code. He wasn't responsible for it. It was completely independent different team. And because it blew up, a lot of people just developed this idea. Ethereum is not secure. The EVM is fundamentally broken. And it's true there are some issues there, but come on. That's like saying, oh, well, Photoshop didn't work. Fuck Bill Gates.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“So you have to evolve in a very methodical way in each step of the way you bring new actors in, they get ready for it. They upgrade their infrastructure to it. And then like shells eventually you get to the outer shell, which is the hard fork for the general public. And if we've done it right, like that blue check mark, they wake up and it's exactly the same. Just get a little update update your client. They start demanding Wagu beef”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“What you do is each step, so the next is white, you go from like 50 people to several hundred, and then purples and open testnet where we want every single person in the entire ecosystem to use it. And it's also a dev net. So that means that people are writing with Plus Playground and local interpreters. Smart contracts can actually start testing them now on the public infrastructure. So it's kind of like releasing dev kits to the Xbox or something like that. You send them out to game developers before you release the Xbox so they can test their video games in anticipation of the release of the system. So you run that for at least a month. And as long as it doesn't blow up in your face, and oh God, what have we done? Hindenburg, all the humanity, you release it and you ship it. The problem is it's no longer in our control. You know, there's over 100 exchanges that have listed Cardano. There's lots of wallet infrastructure. There are thousands of different constituencies. So it's less of a technological problem now and it's more of a coordination problem.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Yeah, so first you start with the Merri era, which is where we're right now multi assets. So you can do metadata and issue tokens on Cardano, but you have limited programmability on-chain. Alonso adds the programmability in, but we already have most of the foundations of the extended UTXO model and the smart contract models. Those things aren't ready. So the first step is, say, fork it. So all those rules are now there. Okay, so that's what we did with Alonso Blue. We forked a testnet-based layer and it successfully survived going from Mary to Alonzo, which means that you can move transactions from one side to the other. Both systems work. And then the next step is to say, okay, well, are the stake pool operators people run the infrastructure able to run this testnet just like they run Cardano? And that's what we're doing right now with Blue. We're bringing all these SPOs in. And then are we able to submit and run smart contracts on the system? And they actually return around trips. You send something, you get something back. Yay, okay. So that's where we're at.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“So we had to name something after church, and it's just weird that we never did. So we said, okay, Alonzo is a good release name. Basically, it's bringing smart contracts to Cardano. It took us a long time to get here, and we'll be there in the next 90 days. Everything we do, there's a process. And so there's all the colors of the rainbow. We start with Alonso Blue and then white and purple. And each step you do some more things. You bring more users in. And then eventually you get to a threshold where you say, okay, everything works the way intended and you push a button. And we initiate what's called a hard for combinator event. And boom, the system has smart contracts. You just wake up and it's there. It's like it's in your house.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Yeah, so Alonzo Church is a famous, famous mathematician, computer science guy, and he was a contemporary of Turing. And there was like these three different views of computing, recursive functions from Girdle and Turing machines from Turing. And Church had Lambda calculus. And they're all equivalent, and they all give you the ability to build a computer. So we like functional programming. So we decided to do that.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Innovation stops the desire to actually do real work stops because you're always inside this virtual economy. So I don't know. It's an interesting question. But drawing it back more to where we're at today, the evolution factors are there. VR is evolving at a very rapid rate. The game engines are just incredible today, and they're really doing amazing things. And there seems to be an overwhelming desire for people to escape the harshness of where they live, just by evidence by how many billions of hours have been spent playing video games. People still play Skyrim.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“But yo, right? You know, I get what you're saying, though. And actually, but then it begs the question and goes back to the very first question you asked in this interview, which is like, how do we know we're not in a simulation? Or, you know, is Bostrom's concepts or these ideas real? Well, it's entirely possible that we are and that we desire to be because the real world is horrifically dystopian or bad, or maybe we actually don't exist. We're completely virtual. And does it matter? And I'd argue that it probably doesn't at some certain point. If you're at the end of your life and your 90s dying of cancer, the fact that you can live out being young, healthy in 25 is probably a desirable thing and no one would ever complain about that. Where it becomes problematic is if the vast majority of society enters this virtual simulacra of reality and as a consequence nothing works because there's no one to do anything. Society falls apart in that respect. There's no desire to do anything in the real world.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Always lurking in the back of your mind that you're in the matrix. You have to not be in the matrix. That's why the bald dude in the Matrix was like, I want to be rich and have a beautiful wife and eat a steak every day. Know, he didn't want to know he was in the Matrix.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Because you know it's not real. But if you could forget that it's not real, then you believe what you're experiencing is real.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“It's hard, you know. But then your knowledge that you live in a virtual world actually becomes a problem for you. So there's going to be this kind of sad dark industry where people try to create amnesia where they're not aware that they're inside the virtual world. And so that's an interesting.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Is just incredible, the latest in real engine. And within one or two more ticks of that clock, the iterations, so five to ten years, the photorealism will be so good that it'll be hard to distinguish between real life and video games. And, you know, the hardware is almost there. So the question is then, when you have photorealistic experiences where you've successfully traversed the uncanny valley to a point where it's good enough, then will virtual reality be more desirable than actual reality. And for the vast majority of people, the answer is probably yes, because actual reality is tough.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Well, yeah, but I guess Homodius, that's kind of like the roadmap there, this hedonistic dystopia where everybody just lives wired into some simulation. And there's some movies about that, ready Player One. And the other one was surrogate and so forth. So, yeah, Hollywood has certainly visualized what this could be. But, you know, I'm not so pessimistic in that.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Just want to have the West world if you can't tell the difference. Does it matter lying uttered to you? Did you ever interview Yval Harari? Was”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Game buying the video game, and then you have this long tail, but you've gotten almost 95% of your value in the first six months. You have a huge churn rate. The odds are the vast majority of people won't be playing the game within 12 months. But if you create an interactive game where there's an actual economy inside the game, then you have YV Online or Second Life or any of these things where you have people playing for 10 years and there's like people buying virtual real estate and all these other things. You as a game developer actually don't have to create a lot of content. So your long tail gets a lot fatter and it generates a lot more revenue and your cost of operating system is fairly fixed or diminishing. So the economics align for doing exactly what you're talking about. I think it'll get done.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“But anyway, the thing is, those assets actually have a blockchain based representation. And so whether the infrastructure that hoists up that game off chain goes on or off, because that ledger exists outside of the game, any person can come in, replicate it, restore it, and turn it back on. SAN's intellectual property. So yeah, it's completely possible to break your architecture up where you have a notion of the player part and then you have a notion of the experience part and you can interchange experiences almost like you do cascading style sheets or something for have different presentations and the ownership of the underlying layer is the players. So yeah, that's definitely doable. And frankly, that's what's going to happen in the gaming world because there's so much value in that. I mean, everybody wants play. And, you know, right now the model is you make a game, you sell licenses, and you have a huge surge of people to begin of the game.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Well, it's basically just a video game that kind of lives on a blockchain, and the creatures within the game can breed with each other and create new crypto kitties, and you can own them. like some sort of dystopian tamagachi with lots of money behind it”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“And then you also have the issue of storage. Where the hell do you put all of it Have exabytes of data? What's the incentive for that?”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“So that's a great thing, and we want to make sure that that's accessible to developers in Cardano. But remember, they're like Bloomberg. It's a centralized speed in that respect. So if you want to build a chain link S competitor, there's other protocols you could do for that. Now you asked about Town Crier, and that was an attempt to kind of sweep the oceans with the net, get the data through a decentralized way. And that was just saying, hey, let's use trusted hardware to go read law, all kinds of websites and other things. And because it's trusted hardware, the scraping is non-biased. If you find something inconvenient to whatever the person who's scraping, the trusted hardware will still do it and it can't be changed. You'd have to manipulate SGX to do that. So that's great, but you still run on the problem of how do you wire that together, the underlying websites still don't have any notion of veracity or reputation behind them.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“A difficult design question. For our part, we love working with Steve and Wolfram, and they're a great company, and they really have some bright people there. And we know on the data set, they're second to none because not only they have it, it's computable. You can do all kinds of things and manipulate with a very rich query language.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Basket, but how do I know that the price feeds I'm looking at are accurate? You'd have to look at Binance and Bitrix and all these other things, or maybe there's conventional Forex exchanges or something like that. Okay, well, how do you weight that and how do you clip outliers? And these types of things, that's a completely different conversation. And there's a lot more mechanics you have to put in for that bundling and attestation of the veracity of the data feed. And what happens if you get it wrong? Your stablecoin gets mispriced and everything goes to hell. And the markets will eventually correct it for arbitrage seeking behavior, but anything that was built on that will fail in the short term. So Oracle is really just a game of, you know, you have to build a standardized interfaces and make it as easy as possible for people to do that and then let people choose how they want to inject data and what level of assurance do they need behind that. And the question is how much do you leave to the user versus how much does the protocol take care of for you?”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“It's a completely centralized thing. It's a video game run by a single company. There's no sense in saying that we're somehow going to decentralize that. What they're just trying to do is extend their currency or NFTs or whatever into new marketplaces. So the minting of that is controlled by a single entity. And the world state of that, you just have to trust Blizzard to inject that into the system. You could try to imagine some sort of like Sentinel group of people within the game who keep Blizzard honest, but it's completely unnecessary because they can change the rules of the system arbitrarily. So in that case, you're optimizing around efficiency and cost reduction. So you'd want a single feed that gets injected into the system from them. If you look at a stablecoin that's algorithmic and it's basing its value on the aggregation of many different exchanges, that's the polar opposite example. Because there you're saying, okay, what's the price of my asset relative to some”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Well, you know, the first step is just say, can we use this as a feed? And they can be what Bloomberg is to the financial markets. So you have a terminal and you have something, and there's always a value of at least offering choice. And so it's not like we're anti-chainlink or we're picking winners and losers. It's an open protocol. It's an open system. So if we're successful, Chainlink will migrate or it will at least support us because they like money. They like users. They like liquidity. It's a disservice to their community not to support a potential customer set. But, you know, you're going to have a spectrum from the desire to do a completely decentralized aggregation, curation, injection, and veracity attestation to a completely centralized vertically integrated set. You need to be able to have that whole spectrum and offer that to the smart contract developer to decide what makes sense. By the way, a lot of cases, they're going to be their own Oracle. So, for example, the world of Warcraft example that I gave.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Because the ones that are tightly coupled with the protocol, they have a built in trust advantage. They've already built a commercial reputation. There's already an increasingly more decentralized set. The other thing is you don't need a token. You can just use Ada. You don't need an Oracle coin for these types of things to work. And by the way, that's just for the injection component and the veracity attestation. So is it true or not? That's not about the aggregation. That's still a tremendously time-intensive, expensive proposition. There's only a few people in the world that have what Steve has with Wolfram. And those guys, by just cutting off those supply to replicate what they have is something that would cause hundreds of millions or billions of dollars. And so it's an interesting question of how do you incentivize decentralized aggregation of information? And that's kind of what Town Crier and other protocols were trying to achieve.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Random numbers for me or whatever. And you're now a service provider. You're making the blockchain full time, but part time, you're doing this. If you're making bagels, you could probably make donuts, that type of a concept. So I think that type of competition is going to be very difficult for a lot of these layer two protocols that aren't tightly coupled with the protocol.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Well, that's basically a real life analogy of the relationship between the Oracle and the smart contract. You're only as good in your infrastruct model as your weakest link, and it doesn't matter if all of your computation is decentralized, if you're at the mercy of your data feed. Because I can just manipulate that and break the entire security model of the system. You'll perfectly execute the wrong answer. So they say, well, if you're trusting us anyway, why don't you pull more of what you're doing on chain into our stack, which creates more transaction fees for them and more value for them. But there are many different ways you can do Oracles. And earlier I was talking about the biology of these things, cell differentiation. The minute that you admit heterogeneity in your system and you start having cells like steak pools or things that are on 24-7, then you can start asking the what-if question. Why don't you guys just also provide data feeds? Why don't you guys also provide state channels or payment channels or generate?”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Running on layer one. Why? Because you have cost reduction and potentially because your trust model collapses to whatever chain link is offering, you're not gaining anything by doing the computation on Ethereum or another platform because you ever watch The Simpsons? There was this beautiful episode where Mr. Burns wants to turn the power off in Springfield and it's the perfect analogy for information security. So he and Smithers, they go through this elaborate series of doors and secret passages and guard dogs and robots and shit to get to the center of the power plant to turn off the power. And when they arrive at the center of the plant, there's like this stray dog that's inside the room. And there's this wicker door that, you know, screen door that leads to the outside. And you're like, well, why the hell did you go through this elaborate series of doors and things if there's like a back door into your system?”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Decentralize it, you can run it on a single device as if it was running on a cryptocurrency. So there seems to be a desire in that community to capture more and more of the smart contract stack and pull more and more of that stack into that layer two infrastructure from”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“About how do you inject and how do you create incentives so that that process over time gets more federated or more decentralized instead of centralizing around one particular setup? Now, a closely related corollary to this is computation off chain. So as I mentioned, smart contracts are intimately connected to an Oracle. The question is how much pre-processing and state management are you going to do outside of the system versus what do you do inside of the system? So it's a very interesting balance between these two. And they were thinking about this stuff for a long time. There's a great paper called Town Crier came out way back in the day at Cornell, and that was all about using like SGX to scrape things. And you can rely on trusted hardware to give you good data. But you could also use those SGX cores to do contract processing because if it runs in trusted hardware, then it's very unlikely to be tampered with or manipulated. And because of that, you don't have to feterate it.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Right, but the only downside is it's centralized. And that's always the Achilles heel of Wolf for him, is he tends to like proprietary things and he tends to like centralizing things and mostly because he likes running the things. Everybody can have an opinion on that. The thing, though, is that after you've done aggregation, there's a question of injection. How do you get that data into the system? And you can do that in a very naive way where you can say, oh, I'm just going to attach a public key to it and it'll sign for that data feed and that injection. And then somehow I'll just trust it as it is. Or you could try to make it more complicated. You could wait data feeds from different sources and have some notion of truthiness or veracity metric or something like that. So Chainlink is just one of many different philosophies that was born out of the academy. I believe Ari Jewell was connected to it and there's some good people on that side. And it has a philosophy about how do you aggregate a philosophy?”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“How many shipwrecks have happened in Florida between 1950 and 2000 that have resulted more than a billion dollars of cargo loss and at least one fatality? And it'll return an answer! He has this incredible source of data that's computable.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Aggregating and doing something. So the Oracle is a super important component in practice for any smart contract involving any notion of value. You need to know when things have happened, how they happen, who won, who lost, etc., etc. So first, where do you get the data from? So what's the aggregator? This is why we love our relationship with Wolfram, because if, you know, one of the things you'll know about Wolfram as you get to know the guy is he's a data pack rat. Every email, every communication, every interaction, he's archived somewhere. Like last time I talked, he said, oh yeah, I have emails from you from 2012. It's like, you still have those? Yeah, every keystroke is written down and stored somewhere. So if you use Wolfram Alpha, it's a similar acrimony of the way his mind thinks. And so you can query the system.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Okay, so anyway, you need the outside world to be ejected into your system, right? Try to keep a straight face. It's great. You need the outside world to make your system useful. It's like all the kinds of things that you care to do with a smart contract usually involve human beings and information streams”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Yeah, I mean, trying to do smart contracts without oracles, like trying to have sex with your pants on. I mean, it's not really fun. It's not exactly the best of things.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Code with the Plutus Pioneers program so you can go to YouTube and watch a lecture and see how that's done. You can do a stablecoin. You can do an Oracle. You can do interactive contracts. It's just, it has to be done a little differently than the way that you would do it in an account style model. Just like you can run an application Java, you can run an application in Haskell. They both can do the same thing, but the code is going to look different and the canonical way of looking at things is different.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Fairly large part of the mind's share of the entire space. So there are no silver bullets, and anytime you pick a particular model, there's an upside and a downside, and there's different ways of doing things from cash register accounting or bank accounting. You can even do different accounting models. But we felt this was kind of the best first step to go into because we started with something very familiar that had a long history behind it, and it maps very beautifully to functional programming principles, this concept of immutability and these things and much more strict management of state and no notion of having this global concept that you have to kind of manage as you break up the system. Now in practice, what does this mean to the developer when they actually start real writing an application? Not too much. There's going to be a little bit of retooling and some new patterns you have to learn, but in practice you can still do the same things. You can implement a Uniswap tile style thing. In fact, we even wrote that.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“System. When you have a concept of this mutable global state in the system, whatever you run locally is not necessarily what you're going to get when you actually push it into the system. So you may misprice things and the contract will fail. It doesn't ever happen in the Pluto's world. So you get a lot of advantages with this particular model. The downside is that it's a little bit less expressive on the boundaries and a little bit harder to write certain types of software with it. But again, how you resolve that is you kind of build higher level languages and other such things that compensate for these types of things and design patterns that compensate for these types of things. The other advantage that we have that's really fun and exciting is that Bitcoin lives in this model and there are other UTXO-based systems and so they're all talking about smart contracts as well and they would like to continue working in the UTXO model. So if you're a Bitcoin contract developer or other things, there's actually already a group of people that understand this very well and that's still a”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Standard mathematician, they'll say, okay, well, is it isomorphic? Is there a mapping between this? What type of function could I actually take something expressed in one structure and transmit it to the other structure and properties are preserved? So we wrote a paper. It's called chimeric ledgers where we actually showed that UTXO, extending UTXO and accounts are somewhat similar in that you can map things that happen in one system to the other system, the properties that are preserved between the two. So in practice, what's nice about extended UTXO is that you can put infrastructure on top of it to make the development experience relatively similar to the development experience of what you would do with Ethereum, but you don't have to worry about this global state. So when you talk about sharding, it's a lot easier to do that. It's a lot more conceivable to do that. And also you get determinism in the system. So when I have a Pluto smart contract, whatever I run locally is exactly what I expect to run in this.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Basically, create something that's still the same as cache register, but now you have programmability, and the big difference is local versus global. So in the case of UTXO, your scripts are your concerns. So whatever's going on in that cache register has no bearing or impact on the other cache registers. But when you look at bank accounting, you have to know the state of the entire banking world to be able to make that work. Why? Because if that transaction is inbound, that wire transfer is inbound. You have to know those funds are actually there. That thing is actually happening. So when you have a global state for a program, it's like you can do a lot more with it, but it's a lot more dangerous. And so you have to build all these mechanisms to try to protect yourself from it. So what we did is we said, okay, add data, add a programmability, and you're kind of in this nice Goldilock zone between what Ethereum did with an account style model and a global state system. And you're not as restrictive as Bitcoin, but you're still turning complete world. You can still run all kinds of things. And then any...”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Tuesday, or you only get this if some event happens like the Broncos when the Super Bowl or something like that. So you need some notion of programmability with it. So a lot of people are trying to figure out in the early days of Bitcoin, how could we improve the expressiveness of the system? And one of ways of doing it is you can go to a different accounting model bank style accounting. So at a bank ledger, every time you do a withdrawal deposit, it's a mutable system. With the cash register accounting, you don't tear up the bills with the bank. You can deduct or add to the ledger all the time. So Ethereum kind of works in that bank accounting system where you send messages, you send transactions, and you're going up or down. And so you can trigger programs the same way. So what we did is we said, okay, if you take the UTXO model and you have some data to it, and instead of saying it's just a digital signature, but it's in a script, you're, you can...”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“So, I guess the easiest way of visualizing it is that UTXL is kind of like cash register accounting. So let's assume you don't have credit cards, you just have cash. And so when you go and buy You pull out your $20 bill, you give it to them unless it comes up to $17.50. They have to make change. So you don't tear your $20 bill and cut a piece of it off and say, here's part of my $20. You give them the entire $20 bill, and then they give you something back and the things that they give you back are also atomic units. They don't cut those things up. So that's kind of what UTXO is all about in a nutshell, is that there's inputs and outputs, your inputs to that 20, and your outputs will be the 1750 that goes to them and then the remaining change that goes back to you. Okay. The problem with this particular model is that the way it was implemented with Bitcoin, there was no notion of how do we run complex predicates, complex contracts on this thing, where instead of just saying, okay, I'm just going to push value to you, I want to put lots of terms and conditions into the movement of that value. Like you only get this if I mow your lawn on.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Right. And also, do you want complete determinism, or is it probabilistic? Because if you relax that requirement a little bit, then suddenly actually you have a broader class of things you can construct this stuff for.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“You don't have to trust the output, you trust the proof, and the proof is deterministic. It tells you these things. So whether you're using zero knowledge, your sigma protocols, or some other mechanism, it's moving you in that particular direction to turn it from a replicated to a distributed problem and go from, I'm doing the work to, I'm checking that the work was done correctly.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Well, Microsoft actually wrote a paper on how to do that. It's called Pinocchio. So that's another example of these types of things, these roll-ups of things, where instead of doing the computation on chain or trying to create some sort of replicated machine that does all this stuff, you instead just say, okay, only thing I'm going to use the blockchain for is to check your proof Going to turn it into a distributed computing problem. Any person in the world can do the problem on any server, untrusted server even.”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“You can do all this computation, and once you've done all of it, you found this magic number that you can verify that the computation was done correctly. So proof of work works this way. Hard to do the proof of work, easy to check the proof of work. Cryptography also works this way. You have some trapdoor where you can verify something's correct, but to get that thing done, if you're doing it brute force, it takes an enormous amount of computation. Well, not all problems are like this, like protein folding to verify the protein is folded correctly, you have to fold the protein. So you have to redo the work. What if for arbitrary computation you could take a problem and then you could generate a proof that you've done that computation correctly and the proof validates in logarithmic time or constant time? Wow, that's incredible, right?”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source
“Brilliance of Alex was to actually realize you could do that and pull those things together, and it may actually have some merit. But by no means is the only guy that does this stuff. There's actually other approaches in verified computing that have explored that. Like my favorite came out of Microsoft research is a project called Pinocchio, and there was a follow-up called Geppetto. And the basic idea was that it's fortuitous that you have these computer science problems like hashing where you can”
2021-06-16 · Lex Fridman Podcast · #192 – Charles Hoskinson: Cardano · IDENTIFIED FROM THE TRANSCRIPT · source