S16 E39: Karl Kreder (Dr. K) on Scaling Proof of Work, Bitcoin & Quai
Search this conversation
0:46Vlad CosteaHello there and welcome to the Bitcoin Takeover Podcast. This is Season 16, Episode 39. We are live with Karl Kreder. I think his last name is pronounced as Kreder or Crater. I'm not sure, but he's better known as Dr. K, the CTO of a project which is called Quai or Quay, whatever you want to pronounce it. It's Q-U-A-I. But he's a very interesting researcher who has been working to make proof of work scale with merged-mined sharded blockchains that are basically interconnected. It's an interesting concept, which sort of reminds me of tree chains, which was a proposal by Peter Todd that goes back all the way to 2012 or something.
1:31But it's more elaborate than that. And I think it's very interesting. That's why we're having this interview, but it was also sort of a vote of popularity on Twitter, where people really wanted to get you on the show. So I really had to make this happen. And before I let you speak, I also have to mention that the sponsors of this episode are LayerTwo Labs, Citrea, SideShift, Bitcoin.com News and NoOnes. Welcome, sir. It's good to have you here.
2:04Karl KrederThanks for having me. I appreciate it. Excited to have a marathon discussion. Maybe we can beat Amir. We'll see how it goes.
2:15Vlad CosteaI'm not sure if you're aware of this, but Amir has been beaten. Like three times since December when he joined the show. So it was, first of all, it was the exiled surfer who did eight hours. And then it was John Carvalho and Eric Voskuil who joined in and were able to speak for eight hours and 21 minutes. So, I mean, if you want to challenge the record that Amir set in December, then I guess you can do that. It's seven hours and 20 minutes, if I'm not mistaken.
2:48Karl KrederSo what you're saying is that we have to go for eight hours and 22 minutes.
2:52Vlad CosteaExactly. Like if you want to beat the absolute record and be number one in terms of endurance, because I don't think there is any more to it, you're going to have to be able to provide information for longer than eight hours and 22 minutes.
3:09Karl KrederFair enough. Well, we'll see how it goes. I'm not as good as monologues as Amir is though. So I'll try to keep it a little more lively, hopefully.
3:20Vlad CosteaRight. So let's start from the beginning. What's your story in the space? When did you discover Bitcoin? When did you start doing any research? I know that you have a background in math and cryptography.
3:32Karl KrederYeah, so I actually got into Bitcoin in 2012 as a hobbyist. And the reason that it was sort of interesting to me was the concept of separating money from the state. So fundamentally, peer-to-peer electronic cash, the libertarian concept of, you know, having a sound money that isn't easily manipulated and controlled. So I was actually doing my PhD during that time. And it was not, obviously, blockchain research wasn't an academic thing, but I was kind of researching myself kind of parallel to my PhD. So I finished that in 2017, and I decided that I wanted to do crypto full-time.
4:20I went to work for ConsenSys. I worked there for about a year before doing a spin-out called GridPlus. I led the development of the Lattice1 hardware wallet. It looks like this. Oh, looks like this. But it's a cool wallet, and it uses, I didn't know I'd get my props ready, but it uses smart cards, which is pretty cool. Or, you know, storing seed phrases. So instead of having to, you know, deal with paper and seed phrases directly, you can abstract a seed phrase so that it just sits on a smart card.
5:05And you can think of it kind of like a prepaid debit card or just really a physically protected backup of your fund. Whereas, you know, if you just have a paper seed phrase, if someone gets that paper seed phrase, they'll have your money. If you have a card-based seed phrase, they need the card plus your pin to get the money. So it's a lot easier for someone to remember a pin than it is to remember a, you know, cryptographically secure 25th word. So fundamentally, I think the card-based solution is how you get self-custodianship in crypto.
5:40And my efforts at GridPlus were, we were the first ones to really do that. So I think it's becoming more ubiquitous. You know, there's people like Tangem now, Coldcard kind of does, or not Coldcard, no, Coldcard, no, it's a cold card. But there's a few companies out there now that are using smart cards.
6:03Vlad CosteaI think it's SATSCARD or something by Coinkite.
6:07Karl KrederYeah, that's right, Coinkite.
6:10Vlad CosteaThere's also Satochip or Satodime. There are different products. There are Satochip, which is sort of like a hardware wallet and Satodime, which is something that lets you use Bitcoin as a bearer asset.
6:24Karl KrederYeah. And you can like pop the resistor out of it to like open it and everything. You know, very interesting. But from my perspective of crypto is always been how do we get it to be electronic cash? Right. And so the effort in using cards to do that is the effort of answering one of the big UIUX problems of crypto, which is how do you get the average population to be able to self-custody it? Because I only think that it's worse with self-custodianship that you get the sort of decentralized properties that are going to be incorruptible by a state entity.
7:02If everybody is using crypto, but they're using it indirectly via proxies like exchanges, you know, I don't think that's going to create a resilient outcome. So we really have to think about with crypto, how do you get it to be used transactionally by a large number of people as possible? And so that's really where I've spent my time is figuring out how to build the tools to allow that to happen. So at GridPlus, I was really focused on the UIUX problem around self-custodianship. And then after that, or as I was doing that, what I really realized is even if you solve that problem fully, you still have the limitations of the networks and the networks, given any sort of popular application, will fall down under use.
7:50So that's when I really started thinking about how to scale networks. And I came up with the idea of how to scale proof of work that was originally called BlockReduce. And that turned into a research project funded by the National Science Foundation that I worked on part time with a professor that I knew at UT. And then eventually in 2022, one of the students that we were doing research with was able to raise money with Blockchain Capital, moved away from GridPlus and started working on it full time.
8:27And then it took us about three years to make the fully produced sort of implementation of a scalable proof of work blockchain. And we launched it in January of this year.
8:41Vlad CosteaOkay, so that's sort of the short story of your involvement in the space. I'm actually curious why you were working on scaling proof of work at ConsenSys, which is ConsenSys with a Y, a company which basically is researching how to improve Ethereum, because I know at that point, ConsenSys was very much concerned with how to migrate Ethereum to proof of stake.
9:09Karl KrederYeah, no, so at that point in time, I was actually spun out from ConsenSys and was working at GridPlus. And the GridPlus was my full time job, and I was just thinking about, you know, how to scale proof of work. And I ended up teaching the first graduate level blockchain course at the University of Texas, Austin. And with that, I knew one of the professors there, and we were talking about some of these ideas of how to scale it. And then out of those discussions, kind of in his research lab, they were able to get a National Science Foundation grant.
9:47And so it was kind of like a part time research project that I was doing while working at GridPlus.
9:54Vlad CosteaRight. So what was it about proof of work that made you decide that this is the way to go, as opposed to following the trend within the industry, which was to find ways to make proof of stake work?
10:10Karl KrederYeah, I mean, fundamentally, I think that if you look at proof of work, it's an economic consensus mechanism, or something like proof of stake is a social consensus mechanism. And I think people that kind of are early to this space. You know, very, very much, even if they can't articulate, like internalize, there's, there's a, there's a complete difference between these two systems. And we can go into the subtleties. But, you know, at a high level, I think it's really economic consensus versus social consensus. And that comes to the fact that in a proof of work system, you can only modify the chain as far back as you can successfully execute a 51% attack, which means at some point, the record becomes immutable.
10:58Whereas in a proof of stake based system, for very little cost, you can mutate anything that's in the chain, as far back as you can coordinate a plurality of stakers. So, right, those, those are different systems. And they also have different sort of characteristics from an economic standpoint. If you look at proof of work, it's a competitive business. Every year, you have to produce new ASICs, you have to figure out how to operate at a lower cost with cheaper energy. Whereas something like proof of stake, you just set up the node, and it's not really competitive business.
11:35So you see the sort of accumulation of wealth going more and more to the original holders of wealth in the system, which creates sort of a plutocratic outcome. Whereas proof of work, I think, is a competitive technical market driven outcome.
11:53Vlad CosteaI also like to think that proof of work is what makes honest money, as opposed to proof of stake, where if you start a network with a proof of stake system, you have to basically mint or issue the entire supply from the beginning. And then the distribution is very tricky.
12:12Karl KrederYeah, no, so there definitely is distribution issues. And then because the system doesn't sort of burn down over time, because of the staking, that distribution is very important, right? I think Ethereum's oftentimes refer to proof of work as as being good for the distribution mechanism, even though even by the time of the distribution mechanism. The beacon chain and proof of stake, I think only 30% of the tokens were actually mined. I believe 70% were still coming out of the Genesis allocations. But even with that, they did, you know, they do pay homage and recognize that proof of work is a good distribution mechanism.
13:05Vlad CosteaAnd there's also the fact that it's very hard to argue that something you created out of thin air, even if the technology behind it is interesting. How are you going to justify the fact that you just printed that money and then you put it on a network and you're saying, well, you have to buy this for $1? You created that out of nothing. There is no expense that was behind it.
13:35Karl KrederYeah, potentially. You know, that certainly is a difference between mined tokens and pre mined tokens. I will say that the concept of fair is like always an interesting one. You know, and people refer to Bitcoin as a fair launch. And I don't know how true I would say that is, because nobody knew really about Bitcoin when it was launched, right? For the first year, if you were just lucky enough to be on the right forums in the right places, you knew about it. So you had this opportunity that nobody else had.
14:12So I don't know if I necessarily, obviously, within the industry, we call something that doesn't have a pre mined fair. But if you actually talk about having the ability to access the system at that time, or know about the system, I don't know if I would necessarily use the word fair to describe that. But I think the more interesting thing, at least from the perspective of Quai we've really been thinking about, is it fair, because fair is such a subjective word is our thinking has always been driven by effectiveness, meaning how do you design the system to quickly grow and grow as quickly as possible, and bring in the greatest amount of capital as possible.
14:52And that's really what drives our thinking about how we design the system.
14:59Vlad CosteaYou guys have been around on mainnet for about eight months now.
15:06Karl KrederYeah, months seven here.
15:09Vlad CosteaIs there any activity in terms of mining? Like, how did it launch? And how is it growing so far?
15:18Karl KrederYeah, so I think our peak hash rate was maybe three giga hash. And then currently, I think our hash rate is probably about 1.3, 1.4. No, sorry, not gig, terahash, sorry. So our peak was three terahash, our current is 1.3, 1.4 terahash. So we're kind of in this accumulation phase presently that you oftentimes see with proof-of-work mined coins when they launch.
15:54Vlad CosteaSo how is Quai being mined? I keep calling it Quay, but it's actually Quai. Does it take a CPU, a GPU, or FPGA ASIC?
16:08Karl KrederSo it takes a GPU to mine Quai. We went through the sort of considerations of all of the possible algorithms. You know, and there's been a lot of ink and breath spilled talking about the pros and the cons. But sort of our thinking was CPUs are sort of the most accessible by the most number of people, but they're also potentially the most easily gamed as well as the most easily bodded. And then there was also the issue of, in our system, because we're this hierarchy of merged-mined chains, we have to do a large number of hash validations quickly.
16:54And with that, something like RandomX, which was really the only known sort of provably CPU-friendly only algo, takes too long to verify a single hash. So that just really got eliminated technically from that perspective. And then looking at GPUs, we like GPUs because they're accessible, they have a low barrier to entry in terms of cost. For a few hundred dollars, you can use them, or if you just have a gaming rig, you can use them. And if you look at the aggregate market of GPUs relative to the supply that goes to mining, it's a relatively small percentage.
17:37So the accessibility to those GPUs is very high because the producers aren't trying to manipulate the distribution to advantage themselves in mining, which is not a true statement about ASIC manufacturers. So a lot of times with ASIC manufacturers, what you'll see is that they'll produce the ASICs and they'll test them for a period of time, really reaping out most of the profits before they actually distribute them to the miners. And the miners had to prepay for the privilege to get first access. So just given the relatively few number of ASIC producers and how that market is 100% soaked up by mining, and therefore wants to be manipulated by those producers, we don't think that's a great situation.
18:25And then the other thing I would say about ASICs is in the event that a crypto system becomes big enough to challenge a nation state, at some point nation states will try to fight back against allowing people access to that system. And if we look at GPUs, they, I'll say this, nation states as they lose control of the money will still be able to control their borders. And with the control of their borders, they'll be able to control imports and exports. So it's much easier for them to limit accessibility to an ASIC by just simply not letting you import it.
19:02But if they tried to do the same thing to a GPU, it would be handicapping AI, it would be handicapping compute, gaming, rendering, really the future of the modern economy. So they wouldn't really be able to limit access to GPUs. So we see GPUs as a more geostrategically resilient hardware to use as the security mechanism for the network. So we ended up picking ProgPoW and anyone that has really a 3000 series or higher can do reasonably well mining a lot.
19:36Vlad CosteaWell, you pointed out to the fact that there is only a handful of ASIC manufacturers, but when it comes to CPUs, not CPUs, GPUs, it's even worse. As you just have Nvidia and ATI, I'm not sure if there's any company that comes close to them in terms of performance. For instance, I know that there was Matrox or something or Matrox, and they were making something that was specialized for industrial use cases.
20:04Karl KrederYeah, no, that's that's a fair statement. I will say two things about it. I think that market as it matures will become more diverse by its nature. We're getting to the point where GPUs are going to start maturing just because of node sizes not being able to decrease in Fed-based systems. And that's probably going to happen by 2030. And with that, you're going to see more and more competitive distribution of production. I think you're going to see people like Intel eventually becoming competitive in that market.
20:37But the real immediate statement about it is Nvidia doesn't limit the supply of GPUs or favor the supply to only certain people because they're using them themselves to mine. And so because mining is only a small percentage of the use of those GPUs, that's I think what makes that market fair. Not necessarily the fact that you're getting it from two different producers, whereas when we look at Bitmain, Bitmain say the only producer of an ASIC, they're going to use that fact to their advantage.
21:12Vlad CosteaI think Nvidia was also making chips for mining, and I think there was also preferential distribution of these.
21:21Karl KrederSo what they did actually is it wasn't a chip specifically for mining. We're talking about say like the 170HX cards. Those were actually de-featured versions of the A100s, which are an inference card. So what they did is they basically burned the, they had like one time set switches or whatever, and they got rid of the sort of floating point functionality on those cards, which you need for inference, which makes them then only useful for mining. So I guess that is somewhat minor specific, but even say the 170HXs still were competitive with the 3000 series cards, certainly at the capital cost point.
22:22Yeah, so fair. I guess the statement is the more producers you have and the smaller fraction of the market for which they're producing to, the better off the system. And we just believe that that works best currently at this point in time in GPUs. I don't think there's really a perfect solution, but we kind of had to pick one. And that's the one we ended up picking.
22:46Vlad CosteaI think earlier I said Nvidia and ATI, and I don't think ATI has been around for like a decade. I'm just old. It's called AMD nowadays. AMD, yep. But yeah, GPU mining, it's quite a hobby. I know that there are a bunch of guys who are using their mining rigs to mine Bitcoin, and then they used it for NiceHash, which is a very interesting business where basically you're mining whatever is profitable, and it's them who directs you to what you should mine, and then you're going to get your payout in Bitcoin. And recently they started running their own AI models with their GPU racks.
23:32Karl KrederYeah, I think AI actually makes the whole landscape for GPU mining better in some ways because it encourages investment in the capital, and it also allows diversification of the use. So maybe crypto mining becomes the last bidder in the AI market, and maybe AI markets pay a little bit better, and then when people don't have workloads, they have excess capacity they can direct towards mining. So from a CapEx perspective, it makes it a very different business model than crypto-only situations. If you look at crypto miners and Bitcoin, the ones that survive are actually the ones not to do the best job mining, but they're the ones that do the best job hedging.
24:12Because if you're mining Bitcoin, you're always perpetually long, and if you build out your CapEx or invest in your infrastructure at the wrong timing of the market cycle, you go bankrupt. So really the people that succeed are the ones that manage their capital correctly because of that intrinsic long that you always have in the market. With something like a GPU, it's a little bit better of a business because you can repurpose that hardware for other use cases. So I think that's also maybe an advantage in some ways for GPUs.
24:41Vlad CosteaI was going to ask you, on the topic of NiceHash, do you believe that the Quai Network right now is NiceHash-able, meaning that someone can rent enough GPUs to 51% attack the network?
24:58Karl KrederI haven't looked at the numbers. I mean, I think we're doing something like 20,000 a day in mining rewards. So my guess is that it could be NiceHash-able, but the way that the node is set up is the sort of depth of ultimate fork is limited inside of the exchange windows. So even if you could do it, you wouldn't be able to double spend any of the exchanges.
25:32Vlad CosteaLet's take a step back and basically allow you to define what the Quai Network is and why you decided to build this and why it's not compatible with Bitcoin and to which extent it's inspired by Bitcoin.
25:48Karl KrederYeah, no, quick question. Can I share a screen on this?
25:54Vlad CosteaI think you can. I'm going to have to make some adjustments.
25:59Karl KrederLet's see. Can you see this?
26:28Vlad CosteaI can. Let me make sure that everything is on the screen. It should be good now.
26:42Karl KrederGreat. OK, so maybe just give background for the inspiration and then the solution here. So again, fundamentally, I got into Bitcoin because I like the idea of sound money and I liked the libertarian concept of separating money from the state. Those are the two statements and I guess maybe the third statement is the way I think you ultimately become resilient enough to do that is becoming electronic cash. I don't think electronic cash is an optional thing. I think you have to have it. So ultimately, we talk about economic resilience, but at the end of the day, the money is ultimately socially resilient.
27:28So this is different than political consensus or social consensus. It's socially resilient and social resilience is really how politically expensive is it to fight the system. So if we look at the last election and we look at Elizabeth Warren and kind of her anti-crypto army, they were kind of fighting against maybe, say, 30 million speculators. And that number kind of was like just enough to sort of bend it that that was a bad idea for them politically. However, if instead of only having crypto and Bitcoin as this sort of speculative instrument class that relatively small percentage of the total population holds and uses, but you had it become the lifeblood of the economy.
28:21You had a soccer mom buy her groceries and fill up her gas tank every day with it. That would, one, get you a much larger number of people and two, because those people are dependent on that system, it makes it politically untenable to attack. And so that's what I see as the ultimate end point that's needed to truly have a decentralized system is it has to be peer-to-peer electronic cash.
28:52Vlad CosteaYou sort of remind me of 2015 Mike Hearn with this statement, maybe Gavin too. They were firm believers that if they make Bitcoin very accepted and adopted everywhere, it's going to be difficult for any nation state to ban it. But the counterpoint to this is that there is a lot of stuff that people use every day and it's still illegal, such as drugs, for example. You can get a lot of people addicted to drugs, it doesn't mean that it's going to pressure the government into not keeping them banned, not waging war against them.
29:28Karl KrederWell, I mean, it eventually does, right? I mean, that's why alcohol is legal, that's why weed has become legalized, mushrooms are becoming legalized in a number of states. That's also a little bit different too, because the power of the money doesn't come from the government, the power of the money comes from the people. It's just we allow ourselves to put value in the money that's issued by the government. The second that we stop doing that, the money becomes worthless. It's not backed by anything, it's not backed by the military, it's not backed by political force, it's backed by your willingness to accept it and my willingness to accept it. Once enough of us say we don't want to use dollars, the dollars become worthless.
30:15That's the fundamental difference here. As system designers, it's incumbent upon us to make a system where we want people in their own best interest to choose to replace the use of dollars with crypto. And if we can do that, you can win the game. But you win the game by creating a better product, which they choose in their own self-interest. So that means it has to be simple enough to use, it has to be cheap enough to use, and they need to use it as a replacement for cash, and then everybody uses it.
30:50Vlad CosteaWell, okay, but why would Quai be what people use?
30:55Karl KrederWell, so it's the question of scale in a decentralized system. So if we can accept that proof of work is the way to accept or create decentralized systems, I would argue that what we've come up with Quai is the way that you enable scale to function like cash globally.
31:18Vlad CosteaWell, but there are other projects that scale, but it doesn't mean that they're going to get the network effect.
31:27Karl KrederI would make the argument that I don't know of any other proof-of-work project that actually scales to function as cash. Just straight out.
31:46Vlad CosteaWell, how do you define scale? Because that's another interesting conversation to have. Like what do you call scalable? Something that enables people to use it at the expense of blockchain bloated, or something which is very lightweight. Because if you want scale, you're going to look at Grin, right? That one has pretty good scalability.
32:12Karl KrederBut it's interactive. So the issues that I would say that you need, and you can't actually get this in a single chain setting, you need to design a system that can replace Visa MasterCard. Because that is like our concept of digital payments. So you need something that has, say, sub-penny fees. You need that sub-penny fee structure to take you north of 50,000 transactions per second. It has to be in a non-interactive environment. And it needs to confirm a transaction in 6 to 8 seconds. Those are the definitions of scale if we want to compete as cash.
32:50So like, what did I do? I looked at what our current digital system is. And I said, how does it work? What are its parameters? Those are the parameters. So if we want to replace that traditional system, those are the parameters we have to meet. So when I talk about scale, that's specifically what I'm referring to.
33:06Vlad CosteaBut wouldn't you agree that this system also needs to be decentralized enough to not have any kind of figure that can make changes to the system? So basically, it needs to have like, a form from which it's just good enough to get adopted without concerns that there's going to be any more meddling?
33:28Karl KrederYeah, no, absolutely. So if you look at this, proof of work is an interesting system. And what I don't think people really understand a lot about proof of work is that it's designed to pit the miners against the holders. And that sort of pitting is what keeps the stasis or ossification of the rules in place. So as the number of miners and the number of holders goes up, it becomes more and more difficult to change the rules. And that's a specific design feature, not a bug of proof of work.
34:05But what that means is that when a proof of work based system launches, it's meant to be ossified. It's not meant to have a roadmap, it's not meant to change. And that's one of the reasons that Bitcoin has trouble really changing is because it was designed not to change, right? Early on, Satoshi and Gavin had social capital that they could use to easily make changes to the code base, as long as they were non-controversial, i.e. they didn't negatively economically affect any subset of participants in the system.
34:37And by the time that we got to the 2017 scaling wars, what it was really saying, because of the length of that debate, was that Bitcoin was really becoming ossified. And it was the last major economic change in the system. And I would actually argue since then, Bitcoin has become less decentralized than it was in 2017. I think that was maybe potentially, or maybe 2018 was sort of peak decentralization of Bitcoin. But that's a design feature. Now, because that's a design feature of these systems, you need the proof-of-work systems to launch, ideally complete.
35:14And so if you look at how long it took us to go from idea of how do you scale a proof-of-work blockchain using a hierarchy of merge-mined chains that allows for sharding of a state without sharding of work, as well as self-organized topologically optimized subnets so that you can get to this 50,000 plus transactions per second. It was almost seven years, right? So we came up with the idea in 2018, and we didn't actually go live with the network until 2025. And the reason that we did that is because we wanted to create a system where we weren't going to be meddling, or if we meddle, we do it very early on.
35:51But we're not spending a decade, kind of like Ethereum does, redrawing what the rules are and what the roadmap is, right? Our goal is to launch the system as it is, fix the bugs or oversights that we may have in the first, you know, 12 to 24 months and tease out. The statement here is if you look at protocols, if protocols are designed correctly, they don't have to change. TCP/IP was created and codified in 1983, right? We are almost 40 years on from that, and there has been three optional backwards compatible updates to TCP/IP.
36:27The only one that you know of is IPv4 to IPv6. The other two are like inconsequential changes. So that proves that even for technology in a rapidly developing world, right, when it was codified, it was still DARPANET. And we have been able to use that all the way through today. I mean, I don't think we're using it on this call. I think this might be UDP, but we use it for every other thing that isn't video streaming, and it works, and we didn't have to change it.
37:00So I guess the statement here is what we believe is that if you design the system correctly and you do it so that it heuristically approaches what is physically possible within the physical constraints, i.e. the lines of the internet, the speed of the processors, and the speed of the disks, you can create a system that is effectively the perfect manifestation of proof of work, which becomes the TCP/IP value. And that's fundamentally what we've tried to approach. And trying to sort of shoehorn that into existing systems and not run from a clean slate is just an impossible task.
37:40Vlad CosteaOkay, I have some questions to ask from this, but it's been 30 minutes and I got to present a couple of ads. I'll start with the one which is actually for something that's merged mined, which is a topic that you like. LayerTwo Labs is creating the drivechain, which is the way to scale Bitcoin with parallel chains that have functionalities that you find in interesting altcoins today. So there's gonna be an Ethereum sidechain, there's gonna be a Zcash sidechain for fungibility, a big block sidechain that's prunable for low fees and a lot of transaction volume, and also something for assets like Namecoin and something for prediction markets.
38:21But you don't have to imagine what this is gonna be like. You can go to layer2labs.com/download right now, and you can get the latest version of the software, which is available for macOS, for Linux, and Windows. And you can try to see how the drivechains launcher works and then connect to different types of sidechains to it, and move funds around, mine some, exchange some from, not from, but to people from the Drivechains Talk Telegram channel. And it's pretty awesome. The only problem is that this needs BIP-300 to get activated on Bitcoin.
38:59I think it can be done with OP_CAT too, but it doesn't seem that's happening anymore. And right now it only works on the Bitcoin signet. But even in this state, I think it's pretty interesting and you should give it a try. Once again, layer2labs.com/download and check it out. And I'm actually curious what you think about drivechains, because you have been researching merged mining for a long time.
39:26Karl KrederYeah. So the statement about drivechains is, and the statement really about scaling, is there needs to be reciprocal validation to get these systems to scale in a way that you can get the money. So you can kind of make a chain, that very intermittently pegs in and out of Bitcoin in a trustless way. But because Bitcoin is unaware of that chain, you can't do it quickly. What we've done in Quai is this hierarchy of merge-mined chains is always reciprocally validated, meaning Prime is validating, in this hierarchy, sort of the most highest chain, is validating the blocks and the lower chains.
40:20And that's kind of the property that allows you to scale without these long times between pegging in and out. It's also the property that allows us to make the system look homogeneous to the user. So in this proposal of drivechains, you just defined three different types of chains that you could potentially have under Bitcoin. And that means that the users have to now still be aware of the infrastructure that they're using. With our system, it has homogeneous security guarantees, it has a homogeneous consensus mechanism, and it has a homogeneous execution environment.
40:58And we also shard the address space so that the user themselves don't actually know or have to know which shard they're coming from or going to. And because of that, it looks as simple as, say, an Ethereum transaction. You give me an address, you give me an amount, and if I'm interacting with a contract, you give me data. And that's all you need to know. I don't need to know if I'm going from this shard to that. I don't need to know if it's ETH on mainchain or ETH on Base or ETH on Arbitrum.
41:30All the system is just a sharded address space with, again, those homogeneous executions, homogeneous guarantees, homogeneous tool sets. So I personally think that if you look at money, a homogeneous implementation is what is required to be able to abstract the technology from the user. If the user has to know that they're on Arbitrum or if they're on base or if they're on a drivechain or if they're on the Bitcoin main chain, you've fundamentally failed.
41:59Vlad CosteaWell, I disagree with that, because when you go to a shop, they ask you, do you want to pay with card or cash? Or card can mean it's going to be Venmo. It can be PayPal. It can be lots of things which are not compatible with the rest of the networks.
42:14Karl KrederThat's a fundamentally different statement, right? The analogous statement would be, do you want to pay in USDC, USDT, TUSD, FUSD? That would be the more analogous statement than the interchange. So that's kind of like the fracturing that takes place. And fundamentally, even the payments infrastructure, if it was just one system, it would like better. You know, obviously, if that system could work for everybody all the time in all the different formats than having many. So, you know, in the same way I go to a store and they won't take my Amex, I'm like, that kind of sucks.
42:55I'd rather just have a card that I could take anywhere, right?
43:01Vlad CosteaYeah, I get that. But my point is that even today, you still have some platforms that you can use for payments, but they're limited by the network effect of that platform. And it's not like you can use PayPal to pay someone who's on the Visa network. You have to basically withdraw from PayPal to the Visa network. So you have to be aware at all times that you're using PayPal. And then you get on the Visa network and then you can turn that into cash or whatever.
43:31Karl KrederYeah, I mean, fair. Like, OK, so in the U.S., there's three networks that have the whole market share. I think, you know, you if you had three systems and likely this will happen within crypto, right? You're not going to have this completely monolithic set. Right. All of these sort of network based systems are going to follow a Pareto outcome, where sort of five networks deal with the majority of the flows. So I don't think you're going to get a complete monolith. But even if we're going to say that five networks are going to maintain all the flows, you're still going to need things that scale far beyond what anything's currently doing.
44:12Vlad CosteaYeah, that's a fair statement. People who are much more idealistic than I am will say that we're going to have everything divided by 21 million Bitcoin. And that implies that we're going to be able to scale to eight billion people. And drivechains is just one way to get there. I'm not saying it's the best possible one. It's the most interesting one that also favors the miners who are the ones who secured the network, because every other L2 that we have right now only indirectly pays the miners.
44:46And this one, you would have the miners being in charge of the side chains and therefore they're going to be incentivized to collect fees from them.
45:00Karl KrederYeah, I mean, there is a way that you could, I mean, if you wanted to actually create a peg in and out of Quai into Bitcoin, or a Quai-like architecture into Bitcoin, you know, you could do that. And depending on how you go about it, there's different ways to go about it, right? So Portal to Bitcoin is actually integrating with Quai. And once they do that, people can take their Bitcoin and make Quai versions of it and get all the properties of Quai with their Bitcoin.
45:33Vlad CosteaI was actually thinking if it's possible to create something like Quai and copy the UTXO set of Bitcoin. So basically everyone who is currently holding Bitcoin can get some free Quai Bitcoin, QBTC or whatever, and also have the benefits of this more scalable network.
45:56Karl KrederThat's an interesting idea. Yeah, I mean, that's potentially something we could do. There are issues around just like laws and KYC, AML and FinCEN around airdrops like that, but potentially there's some way that we could incentivize Bitcoin holders with Quai. You know, that would be something interesting to essentially explore. Now, there is one other aspect to sort of Quai that I think we should discuss and trying to understand. Well, we haven't even really gotten through the technical aspects of how we're doing all this, but there is sort of a monetary aspect that I think is hard to reconcile with something like Bitcoin.
46:51And fundamentally, right, Quai is really built to try to service as a global monetary system that's decentralized. It's not trying to create cryptocurrency. And one thing that we like to look at is history in terms of what has happened in the past. Right. So the United States during the 19th century prior to the Civil War actually operated on hard money. We had a gold, silver, bimetallic standard. And unfortunately, the bankers eventually managed to create the Federal Reserve and create notes. And the reason they were able to do that is there's some fundamental problems with gold, namely that it's hard to divide, it's hard to store, it's hard to transport.
47:37So having paper representations of that gold was more convenient for people, for everyday commerce, and that eventually allowed the bankers to realize they could fractionally reserve it. They did that up until the point that there was a run on the banks, and then that culminated with the Great Depression and in 1933 Act 6102, where Roosevelt confiscated everyone's privately owned gold and put it in Fort Knox. So when we sort of look at systems and we look at Bitcoin as digital gold, we know that it kind of inherits this concept of sound money.
48:14But if Bitcoin doesn't scale, it also inherits some of those deficiencies, maybe not divisibility, but certainly storage and transport. Right. If you can only do five transactions per second, that is a transport problem. On a proof of concept basis, it's really cool because I can get it across the world. And when I first used Bitcoin, it was really cool to see that, OK, I can transmit money around the world and I can kind of get a confirmation in under an hour. But seeing that it's not being able to do that natively to actually run transactions when we live in a modern world where we do tap to pay and you get a receipt in five seconds, Bitcoin really hasn't kept pace.
48:52But it's showing that it has a deficiency relative to transport or sending, right, which is if everyone in the world wants to use Bitcoin, it doesn't work and it doesn't work fast enough. Right, so I guess my fear of Bitcoin is that if you look at where the Bitcoin's going and who's custodianing that Bitcoin and who controls the miners, that you're going to get a similar situation where the bankers are able to co-opt it, potentially fractionally reserve it, which they're already kind of doing, and eventually rug-pulled it.
49:24So that's really like the fear. It's about resiliency of what you get if people self-custody money itself and use it daily. And if you're not using it daily and you're depending on custodians to hold it, you're going to, I think, likely end up in the same situation as what happened to gold. So that's really what we're trying to design against.
49:47Vlad CosteaWell, good luck getting people to self custody, because it's been a struggle since day one just to get people to remember their seed phrases, to remember something as basic as a pin. And a lot of them expect to have some kind of number that they can call to report that they forgot their password so someone can reset that. It's funny. Anyway, we got the best review ever on this livestream from someone called SLCBTC. He said this shit is retarded. Thank you very much. This is comment of the day from the comment section.
50:23I should print that on a T-shirt.
50:27Karl KrederFair enough. But from a point of view, right, like this is a smart card. I found it. Right. So this is a seed phrase and it's protected by a pin. And even if you give this to a 70 year old mother, I kind of use my mom as like the design point for mass adoption. If my mom can understand it, then you've sufficiently created a system that everybody could use. And my mom understands how to use an ATM card or a prepaid debit card with a pin.
50:58And she understands that she loses her pin or she loses a debit card. She's not going to have access to that money. So I think that, you know, that problem is still getting solved. But between smart cards as well as eSIMs on your phones, that that certainly is a solvable problem for the majority of people.
51:20Vlad CosteaYeah. There's also someone who said scale your butthole in the comments. It's very heated. I don't know. I don't know what's happening with this live stream. And there is someone else who's saying Kaspa to the moon and says that you copied Kaspa for the design of the Quai Network. I keep calling it Quay.
51:41Karl KrederWell, maybe maybe we should divert a little bit into the technology. And this is patently different than Kaspa. And what kind of give just an overview of the two features that make it very much different, which is one, it's not an unstructured DAG, which actually allows for smart contracting. It's a structured DAG. So we're EVM compatible and you can get all those benefits. And then the second is Kaspa can scale to some degree, but they're limited in what they can do because they cannot shard. So we can argue about sort of what the saturated throughput of the Kaspa network is.
52:22I would expect it to be somewhere around 2,000 TPS. But to get to global money, you have to go north of 100,000 and we can shard. So we can keep scaling even if a single sort of thread is only doing about 700 TPS, but actually get to global scale money. You have to be able to shard. So those would be my two immediate points with regards to us versus Kaspa.
52:48Vlad CosteaI feel like I've just started some sort of flame war here because I'm interviewing the creator of Kaspa next week. Yonatan? Yeah, he's gonna join the show.
53:03Karl KrederI mean, I'd love to talk, right? So we actually are quite friendly with somebody that was very deep into the Kaspa community until they kind of got kicked out. I don't know like all of the politics around that, but Shai is someone that we work with as a researcher. We had him on a space yesterday and we were discussing Monero and the Qubic attack on Monero. And he's also reviewing some more sort of formal academic papers that we're writing about consensus. So in terms of systems and crypto, just so everyone kind of knows, I love Bitcoin.
53:47I love the idea of Bitcoin. You know, Bitcoin is the greatest asset that was sort of ever created. My issue is I think there are things that we need to do, you know, doesn't necessarily mean Bitcoin doesn't remain a monetary asset. But if we want to remain free individuals in a digital world with digital money, just having Bitcoin does not have all the characteristics that enable that. And I have tremendous respect for the guys at Kaspa and what they've managed to do with a DAG and scaling proof of work.
54:21I don't think that they are correct in their approach, and that's perfectly fine. I have created the system that I think is more correct and has more of the features that are needed to be able to take money to scale and make it so that it's both decentralized, censorship resistant, as well as private. So there's there's a lot of features to Quai that Bitcoin and Kaspa don't offer. But I would also say that in that statement, you know, when we go back and I said there's kind of like a Pareto outcome of distribution.
54:55Right. Maybe we live in the world where there's Bitcoin, Kaspa and Quai. And they're all in the top five. I think that getting proof of work systems to be the top five is really the most important fight that we're looking to have here. I think proof of stake systems just inherently have issues with trust, issues with manipulation, issues with scale and issues with privacy that proof of work systems can solve.
55:23Vlad CosteaI sort of like this renaissance of proof of work. I'm happy that people still work on making it scale, but at the same time.
55:33Karl KrederAnd let me let me make another statement. I own a significant amount of Bitcoin, Kaspa and Quai. Just throwing that out there for everybody.
55:44Vlad CosteaYou spelled out BSQ and you triggered the BSV people in my comments. They're like, oh, well, why do you bother? Because we have BSV and this one scales.
55:54Karl KrederYeah. If we want to dive down to kind of the technical rabbit hole. You know, we can do that now if you want.
56:04Vlad CosteaLet me plug another sponsor, because I do have to do this thoroughly. And then basically I'm going to present them sooner rather than later so that we can have a free-flowing conversation afterwards. Sorry, it's just how this business works. There is also Citrea, which is the first optimistically verified ZK roll up that was ever published on Bitcoin. Basically, what they're trying to build is like Ethereum, like Arbitrum, to be more specific, to enable a suite of financial applications that currently are not available on Bitcoin.
56:43And their end goal at this point is to get some of that sweet liquidity that's locked in on Ethereum as WBTC, which adds up to more than 100000 coins. And it would be awesome if instead of having something which is custodied by BitGo and issued as a token to have something much better, which inherits the security of Bitcoin and also grants you extra options for DeFi and whatever you might be interested in to earn yield on your Bitcoin or lend it or whatever. And they're going to launch this year.
57:20And you can check out the software already at Citrea, which is citrea.xyz. They have a testnet version of their software and you can check it out today. And I do hope it's going to pick up because it's been a while since we had a proper L2, which does something interesting. And Lightning turned out to be sort of a bitter disappointment and I'm keeping my fingers crossed for this one to succeed. So check out Citrea, too. And thank you for sponsoring. And we almost fell down the BSV rabbit hole, which I think technologically is interesting because the idea of having gigabyte blocks in Bitcoin has been around since at least 2011.
58:11With this guy named Dan Kaminsky, who went on stage and did a presentation at one of these developer hacking conferences, which was not related to Bitcoin at the time. And he did the napkin math of what it takes to scale Bitcoin. And basically he used this as a benchmark because he believed that with gigabyte blocks, you're going to be able to match the Visa level of transactions that he measured back in the day. So it is interesting to push this design to its limits, but it's not interesting to try to patent open source software and sue developers and have this wacko who pretends he's Satoshi and cosplays as the creator of Bitcoin 24/7.
58:57So that's my gripe with the whole experiment. What do you think about it?
59:03Karl KrederBSV?
59:05Vlad CosteaI mean, there's BS in it, right?
59:07Karl KrederRight. Yeah, I guess, yeah, we'll just leave that there's BS in it. But from a technical standpoint, just thinking about this as a block size problem completely underestimates what's going on here. The problem that you're actually trying to solve is distributed consensus. And you're also then trying to push data. So the consensus piece is actually like your real limit. It's not. The bandwidth or the throughput. The bandwidth of the throughput is secondary to consensus. So the first problem that you're trying to solve is how do you most efficiently and quickly get to consensus?
59:57And then that provides the bandwidth limit that emerges from the nodes processing time and the wires. And what we've spent a lot of time on at Quai is figuring out how to make the consensus perfect relative to physical reality. So to speak. And just as an aside, I don't know if anybody who is listening knows who Stephen Wolfram is. But Stephen Wolfram is the CEO of Wolfram Alpha, makers of Mathematica. And he's also a physicist, and he launched a project called the Wolfram physics project. And he's an advisor to Quai.
1:00:45And the reason he's involved in Quai is because the network and the consensus mechanism that we came out of in this sort of hierarchy of merge-mined chains creates a macro manifestation of his fundamental theory of physics. So he fundamentally believes that all physical laws are emergent from a causally invariant multiway graph founded by a computationally irreducible process. And what Quai Network is is exactly that. It is a causally invariant multiway graph founded by a computationally irreducible process.
1:01:20But what it fundamentally is, is that to solve distributed consensus, you actually need to model reality. And so when we just talk about big blocks or anything, we are decades past that conversation when it comes to Quai.
1:01:38Vlad CosteaI think it was also Amaury who is like the original big blocker, the guy who created Bitcoin Cash, who said that increasing the block size is sort of like increasing the speed limit on the road. Like just because you allow it to be faster, it doesn't mean that you can physically go faster.
1:01:57Karl KrederYeah, I mean, if you want to create a basic analogy, what we're saying is you want to increase the speed limit. You also want to increase the number of lanes. That is sharding, effectively. And then there's this one other piece which still sort of bounds reality, which is consensus. So if we were to sort of drive the analogy further, air resistance is what bounds your speed on a road. And so maybe what we're trying to do is create consensus in such a way that the road is in a vacuum.
1:02:34So now your speed is really only bounded by sort of how you can go through that vacuum, and the throughput is bounded by how many lanes you have on your vacuum highway, so to speak, if we're trying to make gross analogies here. So the block size does matter, but block size is like 5% of this problem. The other 95% of it is how does data get transmitted, how does data get processed, and how does the network come into consensus?
1:03:06Vlad CosteaOkay, but let me tell you something, because here in the Bitcoin land, and I'm wearing this Bitcoin Army t-shirt, we like to virtue signal a lot. And we say, you can run this on a Raspberry Pi, it doesn't cost you more than a couple hundred dollars to get the hardware to validate your own node and be sovereign, and your node is more powerful than an entire mining farm, because we like to believe so. And also, we take pride in the fact that the average African citizen can afford, assuming that he has internet access, downloads a Bitcoin block every 10 minutes.
1:03:43And this is sort of the metric that's used by a lot of people to basically justify the fact that Bitcoin did not scale, and we never made any effort to either increase the block size or make blocks faster or whatever.
1:03:59Karl KrederSo yeah, I'm sympathetic to this.
1:04:03Vlad CosteaWhat kind of hardware does it take to validate the Quai Network?
1:04:08Karl KrederYeah, so for Quai, we've done test networks that go up to 7000 TPS, and it's the same as an Ethereum validator. So it's eight cores, 16 gigs of RAM, and we say a terabyte, even though right now, you know, you'll probably find 100 gigs for quite some time. The point that I do want to get to in terms of these trade-offs, because I'm sympathetic to these trade-offs, and you know, back in 2012 and 2013 on Bitcointalk, right, everyone was making these arguments about decentralization has to do with how many people can run the system.
1:04:53But as an engineer, we can't make subjective statements, we need to make measurable statements, and we can say, you know, and we can argue over what sort of our driving quantity is. But as an engineer, we at least have to try to assign, you know, measurable properties when we're making a trade-off. So the question that I would pose is, what does Bitcoin actually get from having 20,000 synced nodes?
1:05:26Is that actually more resilient than having 1000? I would also ask the question of how many nodes in the Bitcoin network are systemically important? And I'm not trying to pick on Bitcoin here. This is true of like all blockchains. If you're looking at Ethereum, if you're looking at Bitcoin, if you're looking at Solana, what are the systemically important nodes?
1:05:48Vlad CosteaWell, here in Bitcoin land, we don't call them systemically important, we just call them economic, as in they support some sort of economic activity, ranging from being useful to the miners up until, I guess, being used by a merchant, someone who accepts Bitcoin for payment. But the average person who runs Raspberry Pi in their basement and never uses Bitcoin and they virtue signal about running a node, but at the same time they keep their coins in cold storage, they might as well just keep a wallet on their phone.
1:06:22Karl KrederI would agree with you. I would 100% agree with you.
1:06:25Vlad CosteaRegarding your dichotomy of 10,000 nodes versus 1000 nodes, it doesn't matter the number, it's the geographic distribution that really makes a difference.
1:06:35Karl KrederWell, it's the geographic distribution of the systemically important nodes. I would argue you could take probably 95% of Bitcoin nodes offline right now and nobody would notice. Not a single person would notice. Because the number of systemically important nodes is probably only 5%. The services that a Bitcoin node offers are processing transactions for proposing the next block, which means that service is only important if there's hashrate directed at the proposal, indexing and servicing data to people that are trying to consume it, and ingressing transactions. Those are the three purposes of Bitcoin nodes.
1:07:17That's it. That's the whole thing. If you're not ingressing a significant number of transactions, you're not providing RPC responses to consumers of data, or you're not processing blocks via hashrate, you're not systemically important in the system. There's some concept of the value of having a marginal value of having an additional copy of the ledger on a node, but it's very small.
1:07:48Vlad CosteaThere's also the argument, so I just want to put that out, that running a node increases your privacy, because you're not leaking your XPUB to someone else, and therefore it's going to be stored on your computer, and you can also run it on top of Tor. You can have pretty good privacy in Bitcoin terms, because of course you're not going to have sender privacy, you're not going to have amounts privacy, but network-level privacy and wallet privacy, they're much better when you run your own node.
1:08:19Karl KrederSo I would say that there is a privacy aspect to running your own node, 100% true. However, because of those other things, it doesn't matter. It just doesn't matter.
1:08:32Vlad CosteaIt doesn't matter for the rest of the network, but it can matter for yourself.
1:08:36Karl KrederNo, it doesn't matter for you. The marginal increment you get on privacy, maybe there's some very slight edge cases where you get a certain type of privacy, you get certain people, but saying that you have privacy anywhere in Bitcoin I think is a hard statement to actually back up. And running your own node creates one component, but if any of the other components are missing, I think you're kind of up a creek with terms of privacy. I think saying privacy in Bitcoin for anyone is a little bit dangerous, because everybody that uses Bitcoin should understand that every agency at this point in the world knows pretty much who owns every token and every transaction that's ever happened on the blockchain.
1:09:19Unless you are working with completely virgin mined coins off a block that have never touched an exchange and you execute perfect privacy and you don't ever sort of reveal your identity to where you're spending those coins, everything else is known in the system. It's perfectly open all the way back to Genesis.
1:09:46Vlad CosteaI disagree with that. It's not known, it's assumed. Because you still have some deniability, thanks to the UTXO model.
1:09:52Karl KrederThere's deniability, so like, okay, so if you were using Samurai and if you were using fixed denominations, if you were using proper address hygiene, and if you were running your own node, and you were hiding your IP, and you never KYC'd on an exchange, there's like some deniability in a court that you might have if all of those predicates are met. But there's so few people that have like ever done that. To say that there's privacy in Bitcoin or you can get privacy in Bitcoin simply by running your own node is disingenuous, I think dangerous in terms of letting people think certain things about the system that aren't true.
1:10:32So I would just advise like anyone to, unless you're, you know, like, contributing to open source software, you probably don't have privacy on the Bitcoin blockchain. Like you need a very high degree of OPSEC and diligence to even approach what would be considered privacy.
1:10:52Vlad CosteaBut it's like a superhero cape. If you feel like you can fly, that's all that matters, right?
1:10:59Karl KrederWell, yeah, but then if, you know, the kid jumps off the roof and kills himself, like, I don't want to be responsible for thinking that he had a flying cape on, right? So, you know, as a researcher and as someone who cares about using the system, don't assume you have any privacy in Bitcoin. I guess that'd just be my statement.
1:11:17Vlad CosteaOf course, there's also Kruw, the guy who's running the largest Wasabi wallet coordinator right now. And he's going to argue that if he pays you out of a coin join. So there is this payment that you can make directly from a coin join and you can basically do it after like five or six rounds of mixing. And it's very hard to pinpoint to the source of the funds. And he's going to say, well, I do have some sender privacy because I did send it from this coin join.
1:11:49Karl KrederFair. Again, I would say in Bitcoin, unless you're an OG, like a core contributor, like you still don't have a vote, man. Like you could get it, but like normal people. And it's never going to be perfect. But anyway, I don't know if we need to fall down this rabbit hole right at this time. The point that I'm trying to get to, though, is like everything has a tradeoff. And as an engineer, when we're trying to design the systems, we have to try to objectively weigh tradeoffs.
1:12:22So when we're looking at the network and we're saying, OK, is this beneficial or not? And we're specifically saying how many nodes we have and how many people should be running those. So, OK, there's the systemically important subset of nodes in the network and that's going to be your exchanges. That's going to be your large mining pools. That's going to be your indexing services. Those are the systemically important nodes. Maybe there's a couple from that are like hosted back ends to wallets. Right. So they're ingressing transactions.
1:12:54So that's that's probably a few hundred. I don't know if we want to actually get caught up on the number. We just want to say it's a subset. And more importantly, what we would like to define here is is roughly what is a good node and what is a bad node? And when I think about this, if we can agree on the fact that nodes provide service to the network, we can also hopefully agree on the fact that nodes take resources from the network. So every marginal copy in the network creates every other node in the network to have to provide bandwidth.
1:13:27It has to send the blocks. It has to keep up with the peer-to-peer communication. It increases the number of hops in the network. It increases the latency in the network and decreases the throughput. So these marginal non-systemically important copies, save for the five that are run by false people, that are actually gaining privacy from running these nodes are what I would call leeches in the system, meaning they are taking more resources than they're providing. And so as an engineer, what we say we want to do in a crypto economic system is design a set of incentives such that node runners don't need to be altruistic.
1:14:05And we don't need to tell stories about, you know, Africans doing this or wearing capes and flying. We can simply say that the right number of nodes will come online because they will be able to make money for providing economically important services to the network. And we don't have to think about it as an altruistic process. And if we do that, we can say that the market can decide what the tradeoff is in the resource take versus the resources being provided, and you can get to an optimal outcome.
1:14:36Any time the market decides what it is based off an economic incentive and the market is free and open, what you get is you get a point where supply matches demand. And when supply matches demand, we maximize aggregate economic surplus and maximize aggregate economic surplus. We're saying we're providing the greatest benefit to both the consumer and the producer. And that's what I would argue is always the point of optimality. So it's kind of cool as an engineer. We have to consider the resource. But instead of having to say specifically prescribe how to allocate that resource, how much of that resource you should or shouldn't have, you just get a market incentive.
1:15:12And as long as you have properly designed markets, then clear the market decides the right number of things to have in the system, if that makes sense. So that's really what we think about as why if you're trying to scale a system, you don't want to have as many nodes as possible, especially if they're not systemically important and they're not providing resource because it's a net negative on the ability to create throughput. And you are creating a socialized burden on the systemically important actors.
1:15:44Vlad CosteaMakes sense. I was going to mention something which I guess is relevant not to the topic, but to the context. Another sponsor of the show is Bitcoin.com News. And if you wanted to find out about Quai before it launched, there's actually an article about it on Bitcoin.com in which you guys partnered with Lumerin, which is a business that deals with finding hashrate and basically commodifying it, allowing people to buy miners and be able to monitor them 24/7. I've actually had Ryan Condron on the show a couple of years ago, and I'm familiar with what they're doing.
1:16:30And Bitcoin.com News for those who are not aware, Bitcoin.com has underwent undergone because I said has has undergone this makeover where it's no longer green. They don't promote Bitcoin cash. They don't say that it's faster and more reliable than BTC. Now they're orange if you actually check out their logo. And they followed what the market decided that the real Bitcoin is. And they actually have pretty good news coverage. So it was me who reached out to them and said, you know, you guys have actually changed.
1:17:08And I don't think my listeners are aware of it. And they said, OK, here's some money. Tell people about it. And I'm usually surprised to go to Bitcoin.com News and find stuff about gold and silver and markets around the world that are not the U.S. Because I guess the problem with this whole news space is that it's too U.S. Centric and too much of what's happening in the U.S. Gets reported and takes too much of our attention as opposed to what's happening in the rest of the world.
1:17:41And I like that Bitcoin.com News covers also what's outside of the U.S. And there is also something about Quai if you search for it. And they did not report on your mainnet launch, but they did say some nice words about Quai here.
1:18:01Karl KrederYeah, I know a number of the current and former Bitcoin.com guys. You know, one of the things that I would say with all of these things is we should be a little bit careful about like creating camps and fighting, because the more we in-fight, the more we lose. Right. I don't think anybody that bought Bitcoin in 2012 is the enemy. Honestly, I think that, well, there's a subset of people that are scammers. So like the whole BSV thing, like let's...
1:18:39Vlad CosteaBut they did not buy in 2012.
1:18:43Karl KrederYeah, fair. BCH, like, you know, I think is an honest effort and it kind of was a, you know, a consequence of the scaling wars. I guess my point, like just generically, is that when we think about these systems, we shouldn't think about each other as enemies. The enemy is the government and the bankers, like it's like full stop. So the more that we can have constructive conversations that are less ideologically tied to a specific asset, the better off we'll be, because at some level, right, the people that did Bitcoin.com when it was green, you know, I think they were doing it because they believe that that was better that enables freedom.
1:19:42And I think early Bitcoiners, regardless of any viewpoints, like that is kind of why they got into it to a large degree. So I know a lot of those guys, a lot of them are really cool. I didn't realize I went back to orange. That's interesting. But anyway, I'm sympathetic to all these things.
1:20:04Vlad CosteaYeah. I'm not sure if that statement about bankers and politicians and I'm not sure what else was included in your enumeration, but I'm not sure if this is still popular among today's Bitcoiners. We're like, oh, my God, we have this ETF, we have this politician who's saying nice things about Bitcoin. Oh, look, this institution is finally establishing some sort of strategic Bitcoin reserve or Bitcoin treasury company or whatever. I lost track of what's happening, honestly, and also realized I don't care too much about it. It's just the latest wave of hype and it's going to go away just like the ICO bubble.
1:20:49Karl KrederYeah. Like I say, I think the reason I'm here, the reason I've decided to spend time and effort is I just fundamentally see that we're going into a digital world and there's really two paths. There's a path where digital money and even the crypto systems that are being created open source get co-opted. And maybe they're not a CBDC directly, but they're a CBDC in everything but name. And we will end up in a very controlled world that I think would even keep Orwell up at night. I don't think he could even conceive of what it means if USDT or USDC becomes the next step in money.
1:21:46So there's a second path, which is really in the original ethos of these systems, which is they're meant to enable individual freedom and liberty. And anyone who is in it for that, like you are my friend, let's figure out how to do that, I guess is my point of view. And that's really what I see. I see this as like an existential fight about human freedom. And that's why I've spent time on figuring out how to scale proof of work. And the less that individuals with that intention fight, I think the better off we'll be.
1:22:27And if you look at Quai and the people that we talk to and the people that we partner with, we work with Shai. And he is one of the most prominent Kaspa community guys. There are people on our team that have contributed to Monero. It's not about fighting, it's just about solving the problem. And the problem we're trying to solve is really difficult. And the people that are trying to prevent us from solving that problem are very powerful. So the more that we can find the remaining like-minded people in this ecosystem and work together, I think the better off we're going to be.
1:23:16Vlad CosteaWell, I can agree with this from the perspective that Bitcoin is Linux. And you have all of these distros that do something else or experiment with features without changing the Linux kernel. But at the same time, there is the money aspect that's involved and that makes people become tribal. There's already tribalism in Linux and you're going to have Debian people hating on the NixOS people and whatever.
1:23:44Karl KrederYeah, I guess my statement is, you know, I see how like tribalism can emerge, obviously, and that's why it does emerge. But I guess my point is, if we can talk about problems from an engineering perspective, we can have more fruitful conversations no matter kind of where we're coming from. You know, and maybe some of those just don't get resolved. But ideally, right, we can put our egos aside and say, these are important problems, these are unimportant problems in terms of, you know, alignment and agreement. So just kind of an aside, I guess.
1:24:20Vlad CosteaThere's also the argument that it's much easier to make one project succeed than to make. I don't even know how many coins there are out there. And to some people, it takes way too much effort and attention and time to just find the ones that are legit as opposed to something which is just going to be popular for one cycle and then fade into irrelevance. It's very hard to assess what is interesting and what has people who are genuinely interested in pushing it to succeed, as opposed to something which was only created to make the founders rich.
1:24:59Karl KrederYeah, I mean, those are all problems, for sure. And I can be very sympathetic to that. I'm just, yeah, all of those are caveats, I suppose. But, you know, I still think it's incumbent upon us if we ever kind of want to win this fight, that we indeed figure out how to deal with that. Yeah, from an engineering standpoint, I guess from my perspective, engineering is the mechanism by which you can ideally resolve differences. And the more you engineer to numbers, the more you have conversations to numbers, I think the better off you are.
1:25:49Vlad CosteaI can agree with this to a degree. But still, you're going to have a lot of infighting. It's part of human nature. And people want their fork to succeed. And in regards to what you said about CBDCs emerging without even being called CBDCs, what prevents Quai from becoming a CBDC? Like, does it have any sort of built-in privacy or what prevents this from being co-opted and used as a financial panopticon for all the users?
1:26:22Karl KrederYeah, so again, with Quai, everything we do is trying to optimize for effectiveness. So if you look at something like Monero, Monero is optimizing for privacy. And like I said, we like Monero in some ways, but it's also been delisted from effectively every major exchange. TradeOgre just shut down. You know, that prevents Monero use and adoption. Scale prevents Monero use and adoption. The ring signatures themselves are extremely expensive to optimize. So when we think about systems, we think about how do we get adopted and how do we get used?
1:27:05So in Quai, there's actually two tokens and there's a number of reasons we have two tokens. But if you think of Quai as a global monetary system, a global monetary system actually has three properties. It has a store of value, it has a unit of account and it has a medium of exchange. And so Quai is this EVM compatible token. It's account based. It has a deflationary issuance model. So we can kind of think of that as the store of value in the system. And then the second token we have is something called Qi.
1:27:37Qi we refer to as an energy dollar. And it basically is created by using the proof of work mining process to effectively oracleize the cost of energy. And then we tokenize it and call it Qi. So the idea of Qi is that because it reflects the miner cost of energy over, say, a period of 12 to 18 months, it will be relatively stable in purchasing power. And with that, you create a unit of account which is independent of the dollar that you can use in the system.
1:28:16It's also interesting because unlike Quai or like Bitcoin, because your net expectation is stable purchasing power, not increasing purchasing power, you're more likely to spend it because you don't have sort of the loss of future expected appreciation and value. So having something like Qi in the system encourages velocity in the in-system unit. And then the other thing that Qi has is privacy. So it's built on a UTXO ledger. It uses fixed denominations and the protocol enforces address hygiene. So you can't reuse an address sort of in the same transaction or block.
1:29:01And there's a couple other tricks that we do economically. You can only sort of burn down transactions. You can't put them back up because they're fixed denominations. You have to use pay codes, which is BIP-47. And when you multiply all of these things together, you end up with on-chain privacy. Now, to your point, though, you also want node privacy of ingressing those transactions. And we have other tricks for that. But the idea is that Qi has cash like privacy properties with stable purchasing power is the goal of that token.
1:29:36And so between the two tokens, what we're trying to do is create a monetary system where you both have hoarding as well as velocity. And it's only with velocity that we believe that you get sustainable economics in a mined system over time. Right. So if we look at Bitcoin and sort of the security margin that's going to miners right now, that is pretty much fully being paid by inflationary subsidy. And at 5 TPS, you know, as that inflationary subsidy continues to decrease, transactions have become very expensive to make up for the mining security budget.
1:30:16So kind of at Quai, what we're saying is, well, instead of having 5 TPS, let's have 50,000, and then instead of having a very expensive fee, maybe I can have, you know, this relatively low, you know, one penny fee and still generate enough revenue. To pay for enough mining to make it secure and decentralized. So that's kind of the economic design of Quai. When we look at kind of, so there's this thing called Gresham's law, right, which is good money pushes out bad, or sorry, bad money pushes out good.
1:30:51So how this manifests in Bitcoin. I think it's the other way around. Yeah, no, so bad money pushes out the circulation of good money. Right, so if I have Bitcoin and I have dollars, I'm going to send you dollars before I send you Bitcoin. It's only in a very special situation where you're on some other country that I don't have good banking access to where I'm going to pop open the Bitcoin wallet and send you Bitcoin. It's because I don't have dollar rails. But in any other situation, I'm going to send you dollars if I have dollars before I send the Bitcoin.
1:31:27So that means that dollars have naturally more velocity than Bitcoin and Bitcoin is hoarded. And you can see that both, you know, how wallets are used, how Bitcoin is used, where it's held, how long it's held for, et cetera, et cetera. People have a preference to move dollars. So even if Bitcoin could instantaneously scale, the fact that it itself is a hoarded quantity, not a spent quantity, would only be getting people to use it through some sort of altruistic concept. But the natural behavior of people will be to hoard it, not to spend it.
1:32:02So kind of the idea with Quai then is we have the two tokens, one which you want to hoard and Qi, which you're fine spending because it doesn't have increasing future expected value. And it also simultaneously gives you those privacy properties.
1:32:17Vlad CosteaThis sort of reminds me of NEO, which I think had the NEO token as well as the gas token.
1:32:27Karl KrederSo gas in this system is paid with the respective token. So if you do a transaction in Quai, you pay it with Quai. If you do it in Qi, you pay it in Qi. Really, the difference is you want an account based system that does smart contracting or you want a UTXO with potentially increasing future expected value, or you want a UTXO based system with stable purchasing power and cash like privacy. So that's why you pick one over the other.
1:32:55Vlad CosteaLet me tell you a user review of our conversation. Not user, a listener's review. He said, and his name is Kano or Kano, everything that comes out of this guy's mouth is a dumbed down version of what Craig Wright says. This is the worst kind of insult, man. Like just just call us idiots or something like normal people.
1:33:22Karl KrederYeah, I mean, yeah. OK. Yeah, if he wants to specifically refer to something, I would love to hear what he thinks a bad argument is and what his counter argument is. I think Craig Wright is an idiot and everything he says is retarded, but cool.
1:33:45Vlad CosteaHe has more money than your country. But other than this, I do want to branch out from your previous statement and ask you basically why is it such a big deal? It's such a big deal to have EVM compatibility because you have basically two schools of thought. You have the ones who say, well, it's the standard by now because it's been around for 10 years. You can port smart contracts and it's so easy to use. But on the other hand, you have the critics who will say, well, it uses a programming language, which is horrible.
1:34:16It basically is not secure enough and it should be more like Rust and less like JavaScript when you program smart contracts. And why did you decide to use this as opposed to something else which might be more secure?
1:34:38Karl KrederYeah, I mean, great question. Ultimately, again, I guess the compromise is the question between sort of how do you maximize adoption? And I would certainly agree with the criticisms of the EVM that it's not well designed necessarily or it could be better designed and that it's relatively slow. Now, obviously, you can use access lists to get parallelization, which sort of ameliorates a lot of those effects. And then you can say the programming language, the programming language, actually, in terms of JavaScript is upstream of the opcodes.
1:35:13So you could write Rust to compile into bytecode very easily. So the Rust versus Solidity argument isn't necessarily like a good one. So it's really just what ecosystem exists. What do you think you can get the best adoption across from? And we chose the EVM for that reason. In terms of the VM choice, we see that as almost one of the least consequential choices in the system. If you look at Quai, you've already mentioned that it has two state roots. You have a UTXO root and you have an account root.
1:35:53And it's actually very easily potentially modified at a third state root, which would be a relatively non-controversial change because the total amount of processing or gas in a single transaction would be fixed. And the interoperability between the state trees would look just like it does between the existing state trees. So if, hypothetically, you could add, say, a Wasm or a Move VM into the system, and it wouldn't really have much added technical complexity or cost. But we don't think the arguments for having other VMs exist.
1:36:35The programming language argument isn't really a good one. The efficiency argument potentially is there, but we also don't see that as a major bottleneck given good parallelization and processing, I guess would be the statements. So you don't need Rust.
1:36:57Vlad CosteaWell, to be fair, the two people who, actually it was three, the three people who came on the show to criticize the EVM are building alternatives to it. So there are the Alephium people who are building something which they call closer to Rust and much more robust for building smart contracts. And there's also Andrew Poelstra from Blockstream who recently launched Simplicity, which is a programming language for smart contracts, which for now is only available under Liquid Federation. And they hope one day it's going to come to Bitcoin too for scripting.
1:37:34Karl KrederYeah, like I said, in terms of the actual language people program in, most things are compilable. Now, like I said, if we had infinite resources to make sure we're not adding surface area, adding a Wasm VM would be a great idea because Wasm already supports many different programming language. It just doesn't have the breadth, depth, and auditable support of existing contracts. And I think that's probably the reason we started with the EVM. But yeah, I'm certainly not going to defend that one tooth and nail. It wasn't a great design or implementation.
1:38:19Vlad CosteaWell, it's the one that became popular for some reason. And I also asked Vitalik when he was on the show why he does believe that of all the Bitcoin 2.0 alternatives that were out there from Counterparty to Mastercoin, there was also something called NXT, there was BitShares. It's not like Ethereum was the first or the only one of its kind, but for some reason it's the one that succeeded. It's the one that we still talk about today. It's the one that set out a standard for how to build smart contracts and applications.
1:38:57And right now, even in Bitcoin land, you have projects like Citrea, which is acknowledging the supremacy of the EVM by trying to bring it to Bitcoin on a roll-up. And you have the drivechains, which are also trying to bring an Ethereum sidechain to Bitcoin. So it's almost like there's a general acknowledgement that this is the one that got the network effect and the one that people know how to use and to build on top of it.
1:39:30Karl KrederYeah, I would agree with that. I mean, that's probably the fundamental driver of the choice for us. But we do recognize it does have shortcomings. And like I said, given the construction of Quai, adding a better VM in the future, given the fact that we already have multiple state roots, would not be, I don't think, difficult or controversial in terms of the system.
1:39:57Vlad CosteaSo now that I saw the news from Bitcoin.com about Quai, I gotta ask you, are you guys into mining Bitcoin? Or are you affiliated with a company that is involved in mining Bitcoin?
1:40:13Karl KrederAs an organization, I mean, we have partnerships, right, with Lumerin, with Portal to Bitcoin. There's cloud mining partnerships, but we're not like directly affiliated with a Bitcoin miner.
1:40:28Vlad CosteaYeah. It is interesting because you guys like Proof of Work so much.
1:40:37Karl KrederYeah, I mean, you know, we kind of, there's not many Proof of Work projects that are out there, right? I mean, you've probably mentioned all of them already.
1:40:49Vlad CosteaThere's also Ethereum Classic. I think this one doesn't get enough love. It was only cool for as long as there was hype behind it. And I guess Barry Silbert really liked Ethereum Classic and was hoping that it's going to be the one that will take all the miners once ETH goes Proof of Stake. That did not happen as expected. There was even a fork of ETH that had more hashrate than Ethereum Classic at one point.
1:41:21Karl KrederYeah, no, fair enough. Let me share my screen again. Maybe we can kind of give a view of kind of what Quai looks like. And I'll do it kind of in steps here. This might be the easiest way to kind of see it. So this is actually like mainnet running. So this is kind of the hierarchy of chains. You can see this sort of purple shares or purple blocks we call a prime block. Then we have region block. Under that we have zone blocks. And then these things below it are actually called workshares.
1:42:00So this is the current sort of state of the network. The network is sharded, but it uses the minimum number of shards based off of the current demanded TPS. So we're only doing about 7000 transactions per hour, which means we only need one shard, obviously. But the system can measure its own performance based off the uncle rate. And then once the uncle rate sort of achieves a sustained level over 20 percent, the protocol allows a new shard to get mined. And then it goes into a multishard setting.
1:42:38If you look at what a multishard looks like. Let me go over to that. So this would be. I mean, here it comes. This is what Quai looks like in what we would call it a two by two setting. Here, let me play with this. So so that means this has four shards running total. You know, those workshares are happening below the zone blocks, but the system we've done testnets actually with nine shards, and like all the testnets with nine shards. We've done private testnets with up to 25 shards, but the system itself is limited to 255 shards. So this is kind of the mechanism by which you can get scale and proof of work.
1:43:38So you can get fast transaction confirmations. You can shard state without sharding work. It kind of pushes the limits of what a browser can render. That's why we only did the two by two demo. It might work better if I refresh it. See here.
1:44:02Vlad CosteaWhat's up with the music?
1:44:05Karl KrederOh, you like the music? It's a little. We can do a little Darude, man.
1:44:12Vlad CosteaThis is like some sort of RTS from the 1990s, like retro polygons.
1:44:19Karl KrederI think I think Darude was published in a... Are you saying the music or the visualizer?
1:44:27Vlad CosteaNo, the visualizer.
1:44:29Karl KrederThis is being rendered like using a library in your browser. So there's there's only so much you can do with that, but it is showing you the structure and above of the chain. Again, this this is a two by two. So in this structure, the system can handle. We estimate somewhere between 2,400 and 3,000 TPS, but you can keep sharding it out to 255 shards. So at that point, we think it'll be about a quarter million TPS that you can push through.
1:45:07Vlad CosteaYeah. So I know that when it comes to sharding, the biggest challenge is to coordinate the shards to make sure that they're in sync. And they're able basically to write and send out the correct information to the other one. If you take down a shard, it can create issues with synchronizing the whole chain structure. And how did you manage to coordinate this complex?
1:45:37Karl KrederThat's actually that's actually like where, you know, we spent a lot of time. We have a novel consensus mechanism. It's called PoEM or Proof of Entropy Minima. So basically what it does is it allows us to never have ties. And another statement to say is it's deterministic or it's causally invariant. So instead of measuring, just sort of giving the weight of the difficulty to a block, we actually preference blocks by the actual achieved difficulty. So in things like early Bitcointalk, people might have referred to this as chain work.
1:46:27So we kind of measure it like chain work. So instead of thinking about having a block that, say, has 60 leading zeros, we basically take a binary log of the hash and we can then weight a block and say this block has reduced 63.268495 bits of entropy from the system. And then that weight is as unique as a hash. And so in this system, when you look at these zone blocks, they're actually all leading the system simultaneously. So it's not technically a hierarchy. It's technically a lower hierarchy.
1:47:02But because in this multilevel system, you're guaranteed that you never have a collision of preference. You can shard really without limit to the number of shards you can have. So basically, in each layer here, the weight of the sub-shares rolls up into the dominant shares. So they inherit the entropy reduction for their block weight. And even that, again, is deterministic. But then what it's really saying is that the weight at any point is the summation of all of the entropy that's been reduced out of the system up until that point.
1:47:44But it gives us fair comparisons of shares, of blocks, of super blocks. And it makes sure that you're never going to have a colliding preference and choice within the network. So as long as people are mining to the incentive, they always agree with each other. In Bitcoin, we kind of have two blocks, say I have block A and I have block B. We have time that it takes for the network to figure out if they want to mine A or B based off who builds the next block, C.
1:48:17So in expectation, that's 10 minutes. In this system, if I have seen both A and B, A or B will necessarily be preferable to everybody. Now, with chain work, the problem is you create withholding attacks. So basic chain work treatment of this isn't correct either. So what we actually do, and the reason we call it entropy, or entropy minimization, is because what you're actually doing in the system is every time you begin to mine a block, there's a certain number of possibilities of the outcome. And if you're using a 256-bit hash, the possibilities of the hash outcome are 2 to the 256.
1:48:56And as I come up with block proposals, I'm shrinking the number of future paths in the system. So if I have this big cone of possibilities, every time I find a hash, I'm shrinking it down, down, and down. And eventually, I've shrank it enough that we say we have a block produced and we can move forward in space. But we're basically saying we prefer the ones that shrink it the most. And then that's why it's entropy, because that's just the Shannon entropy treatment of information systems. And that also, in some ways, is chain work, if chain work were defined very specifically.
1:49:30Now, the next piece of this is that you can get a withholding attack if you simply use chain work. But the thing that Satoshi—well, it wasn't even really Satoshi, because he did the longest chain, not the heaviest chain—but the thing that Bitcoin doesn't do is it doesn't treat blocks as a dependent construction. It treats the weights as independent events. So when you meet the threshold, you're basically just saying, I assign the threshold and I'm adding up the thresholds over time. And that works in a strictly linear chain.
1:49:59But the right way to do it is to understand that if this block is built after this, block two is built after block one, block two points back to block one, two is a dependent event on block one. So instead of sort of adding in the linear field, you should be multiplying, which is adding in the logarithmic field, which then allows you to do chain work without creating a fold. So this structure uses PoEM to adjudicate who the leader is, which is a concept of chain work, but with a combinatorially correct treatment of how that work adds up over time.
1:50:39Vlad CosteaSo just for the record, what we are seeing is transactions or users?
1:50:46Karl KrederSo this is not the live network. This is a simulation of the network if it had four shards running. The actual network, because we're only doing 7,000 transactions per hour right now, is a single shard, and it looks like this. This is when the network expands and actually starts sharding, it'll look like this. So what you're seeing here, to answer your question, the blue blocks are what we call workshares, the green blocks are zone blocks, the yellow blocks are what we call region block, and the red blocks are what we call prime blocks. So if we think about the structure of the system, you have a hierarchy of merge-mined chains. So every miner, when they're mining, they're proposing a header that has headers that has prime chain content, region, and a zone. So they get to pick what shard they're mining in, but they're always proposing a header that could potentially be valid in all three.
1:51:57And then based off the difficulty they achieve when they mine a block, they'll get either a zone block, a zone and region block, or a zone, region, and prime block. These are referred to as either super blocks or coincident blocks. But what that does is it creates blocks which have a shared hash across multiple chains in the hierarchy, and what we do then is that allows for a hash link referencing between the shards. So when you're talking about how does the system coordinate, it coordinates by the production of those dominant or super blocks when you're mining, and the super blocks are what coordinates the system.
1:52:39So the cool part about the system is not only are we sharding, but the mechanism of coordinating the shards is the same security and incentive mechanism as just mining a single shard. So what is the consensus mechanism? It is mine the longest chain, and then when you're in a multishard setting it's just mine the longest shard. And so we don't have any sort of secondary consensus or math or anything funny, we're just using proof of work with a different mathematical weighting system we call PoEM to allow that emergent structure to exist.
1:53:16Vlad CosteaRight, so from the point of view of the miner, can they actually choose which shard they're going to be mining? So basically you can have some shards which are over secured and others that are under secured?
1:53:33Karl KrederSo the system makes it, it adjusts the rewards to equilibrate the hash. So if there's too much hash in one shard relative to the other, it drops the reward in that hash and raises the reward in the other until the hash rate is equivalent in all shards.
1:53:54Vlad CosteaWell that's an interesting approach.
1:53:59Karl KrederSo it self-balances, right, because we can measure the hash rate, we can measure the hash rate cross shard because we have the coordinating layer of the higher chains, and then the difficulties adjust until they all match. Or sorry, the reward per hash adjusts to get the miners to switch. So you can mine more than one shard and the other, but if one shard has twice the amount of hash as the other shards, you're going to be getting half as much pie per block.
1:54:28Vlad CosteaWhat about the users? Are they aware that they're using, or that they're on a certain shard for their transactions, or is it automatic?
1:54:41Karl KrederThat was like an important design criteria for us, is that we didn't want... I like that song though, actually.
1:54:54Vlad CosteaLet's not get removed from YouTube.
1:54:57Karl KrederYeah, right. But yeah, so that was like an important design criteria for us, which is we don't think sharding is really compatible with scale. Or sorry, making the user aware of sharding is compatible with scale. So one of the things that I find frustrating is even if, say you're... I personally think like the Ethereum ecosystem is a great example of sharding, right? When we look at the L1-L2 construction, the L2s are fundamentally like shards of the L1. So when we wanted to, say, a transaction, we don't just need to agree on the asset, we also have to agree on the shard.
1:55:47And the way that you do that in Ethereum, you say, okay, I'm willing to accept USDC. Okay, are you willing to accept that on mainchain? Are you willing to accept that on base? Are you willing to accept that on Tron? Are you willing to accept it on Arbitrum? And every time you deal with a contractor, you kind of have this back and forth of acceptance. Because you're like, okay, I have USDC on ETH. And they're like, well, I want it on Tron. So then you're like, okay, I guess I'm gonna have to...
1:56:11And you're like, well, I don't have any on Tron, do you have any on base? Right? So you kind of go back and forth in that specification to try to resolve how you're actually going to send assets to each other. And this system, it's a sharded address space. So the addresses in a given shard, the prefix actually determines where that address is valid. So currently on mainnet, all the addresses start with 00 because 00 is the first shard. When it expands out to the next shard, the valid prefixes will be 01.
1:56:46So they're encoded in the addresses and the wallets will just generate shards, but the users don't have to actually think of it. So if you just give me your address, regardless what shard you're in, your address encodes the shard. So then the users don't have to worry about where it's coming from or where it's going to. They just need an address, an amount, and they can send money throughout the system.
1:57:10Vlad CosteaOkay, interesting. Can this potentially be bad for privacy? Because you know that the address comes from this shard and therefore the user is using that one.
1:57:29Karl KrederSo in the EVM side of this world, we don't really think of this as having privacy. In the UTXO land, you are correct in that you are subsetting the aggregate system into the number of shards that exist. That's what you're saying. Now, I'm not revealing a lot because if I have four shards, I'm basically subsetting the system into four buckets. So, okay, I kind of know that when you're transacting, you're transacting in bucket one, not two, three or four. But your anonymity set would still be the size of the bucket.
1:58:10Vlad CosteaCould you have like a fungibility problem where the assets on one shard are more valuable than the ones on the other shard?
1:58:22Karl KrederYou could, but the system is designed to try to prevent that. So we prevent that by adjusting the rewards. We also prevent that by making the assets very simply interoperable. So the trade-off in scaling the system isn't settlement time. It is the time to re-spend money that is sent across shard. And that time goes up proportional to the number of shards that exist. That makes sense. So if I send you money and we're intra-shard, we get Visa MasterCard confirmation and settlement. Okay, so, you know, this is like five to eight seconds.
1:59:09If you're in a different shard, and let's say, and we're in this four shard topology, which is say this 3,000 TPS, that confirmation time is still five to eight seconds. The settlement, meaning the time which it would take before you can re-spend the money I sent you would be on the order of 40 seconds. If we go to a system where there's, say, a hundred shards in the system, and Quai is sort of ubiquitously used worldwide as the only financial layer in existence, that time turns into something like 90 minutes.
1:59:49And that's the trade-off in this system.
1:59:54Vlad CosteaI'm happy that we discussed this, because everything sounds too good to be true, right? But there's always a trade-off.
2:00:03Karl KrederYeah, I'm a perfectly open book. So I'll tell you anything about the system and I'll be forward about what we're doing and how we're doing it. What I would argue is I don't think it's that impactful of a trade-off. So when we look at the current money system, it's all Visa, MasterCard-based. And if I'm a merchant, I don't actually get access to that money for goods that I've sold. For 60 days, I don't get any of it. And I don't get all of it for 180.
2:00:34So presently, the world somehow operates with merchants not actually receiving their electronic payments in total for 180 days. And what we're saying is if we replaced every other system in the world, that time would shrink down to 90 minutes, worst case. And I think there's very few businesses which are so capital-efficient that that actually is a meaningful trade-off. I think you're just improving on everything. I don't think that there's that many capital-efficient businesses. Now, if there are, there's ways to get around that, which is try to organize your commerce so that it happens intra-shard, not inter-shard.
2:01:15And you can actually also do atomic swaps cross-shard, if you have a liquidity provider that's willing to front the cap. So in any case, right, it's very simple to cross shard, you're only going to pay like a couple cents to do it. And the time that you're going to wait is, you know, depending on the system, is anywhere from a minute or two up to an hour and a half. If again, it is sort of the financial infrastructure for the entire world.
2:01:46Vlad CosteaWell, that's a very big statement. The financial settlement layer for the entire world.
2:01:53Karl KrederYeah, no, I'm saying like that to get to that 90-minute delay to settlement, it would have to be the financial infrastructure for the entire world, just in terms of like throughput and the number of shards. So to get to that number of shards, basically all economically relevant human activity would be happening on the chain.
2:02:17Vlad CosteaI mean, I do like the research behind this and I do like the energy, the optimism. It's just that when I think about it, you see that even today Ethereum is probably close to the all-time high, but it's not near the all-time high in terms of usage. So the price went up. There is a lot of activity with ETFs and Wall Street getting in, but it seems like the spirit is fading just like it did with Bitcoin and the users are not there to sustain the growth.
2:02:51And basically you're doing the reverse. You're trying to provide something that people can use to build a parallel economy, to create their own parallel systems. But it's 2025 and you launched in January, right? And the question is, does the world still care about this or have we moved on to something which is purely speculative and then we just pretend that we care about the technology and we want to replace the dollar or whatever. But the technology is just there in the background to justify the money printing machine and the fact that there is a lot of speculation here.
2:03:34Karl KrederI agree with a lot of your statements and I think that's what we're fighting. Ideally, we would have figured out seven years of research and three years of full-time work in six months and launched in 2018. Unfortunately, and we weren't like the only ones trying to do this. I don't think anyone has gotten to a system like this. I mean, even Kaspa, it took them time. So I fully agree with this sentiment at some level. It's just when the rubber meets the road and you're actually trying to build a decentralized system that can scale, it just took time to solve. So unfortunately, we couldn't get it done faster. But yeah, that's the idea right there. And if we're to compare this to Kaspa, the difference is that each one of these shards should be able to do, roughly speaking, what Kaspa can do, but it can also do it with smart contracts.
2:04:41And we can have more shards as the demand in the system requires them. So like I say, I think with Kaspa as a researcher who kind of tries to understand how the systems work in terms of latency consensus and data propagation, there's a limit to where Kaspa can go. And our limit is potentially 255 times larger than that, I guess is the statement. And so with Kaspa, it's like they've certainly made improvements. They've certainly made innovations. It's just without sharding, you can't still get to money. You can get to replacing one interchange in one country.
2:05:26You can't get the money.
2:05:30Vlad CosteaWell, I think that they're using pruning, which is another way to basically instead of storing everything, they're going to drop the old data and avoid the bloat this way. You guys are using shards.
2:05:47Karl KrederNo, it's not about the bloat. It's about consensus. So the bloat, you can do techniques to get rid of it. You can impute state rent on people. You can prune dust out of chains. You can do archival versus light nodes, etc, etc. Those are just techniques for dealing with the data that's accumulating on chain, but that's not what's defining the throughput. What's defining the throughput of the chain is how quickly you can get to consensus. So by isolating consensus such that we can create these sharded local views of consensus prior to having to get to this global view of consensus, that's how you're getting increased scale.
2:06:33And so their limit isn't the bloat. It's the time that it takes them to get to consensus in a single shard. That defines the bandwidth. And so what we're basically saying is that we should have the same bandwidth as them in a single shard. But the trick we can do that they can't do is we can create a new shard. That's like the trick. So we're under the same physical constraint of getting to consensus in this single shard as they are, but they don't have a mechanism for adding a second shard, we do.
2:07:09Vlad CosteaFair. I just want to announce that we surpassed 1000 views on livestream. Actually, right now it's at just a second, almost 1200. Not that bad, considering that this was basically we we arranged this interview on a short notice. We are supposed to do it next week. But the interview that I was supposed to have this week got postponed and I asked you if you can join in and it's happening. We're here, we're live. We're talking about Quai and also have to present a couple more sponsors and then we can move on to the next point from our agenda.
2:07:55First of all, I'll start with SideShift, which is the place where I tell people to go to swap their stablecoins for free market money. Because why would you hold on to an asset which can be censored and has this central planner that can basically go down overnight and leave you holding pointless tokens? So if somebody offers to pay you in stablecoins, which is very popular nowadays, it's happening. Unfortunately, Bitcoin did not get the network effect that we wanted for payments. And as you can also find out from the previous episode of the show, which featured Bitrefill people who are selling gift cards, you're going to find out that Bitcoin is not as popular for commerce as it used to be.
2:08:43And stablecoins are slowly but steadily replacing Bitcoin as the preferred way to make digital payments across national borders. And my advice for you is that we're in the middle of a bull market. You should not be holding on to stablecoins. Just buy something, not buy, just hoard something which makes more sense to you. I just checked and there is no Quai on SideShift, at least not yet. But there is Bitcoin, there is whatever else you think is interesting, but they don't have the privacy coins, unfortunately.
2:09:21Maybe for regulation purposes and also SideShift, as far as I know, doesn't work in the US. But I guess if you're smart enough to know how to watch porn without having to put your ID online, you can also use SideShift, not financial advice, not Internet advice, not anything. Anyway, the website, the URL is SideShift.ai and check it out. It's pretty cool for swapping. They have a lot of liquidity and they also have pretty low fees. And last but not least, during this episode, there is NoOnes, which is a peer-to-peer marketplace which aims to empower the Global South.
2:10:07The CEO of this company is Ray Youssef, who is probably known in the history of Bitcoin for creating Paxful and running that for many years. And now he decided that he doesn't want to deal with VCs from the US. He doesn't want to deal with suits anymore. He wants to build something with the sole mission of empowering people from Africa, from Latin America, from the Middle East, to have access to this parallel economy and be able to increase their standard of living by becoming hustlers on his platform.
2:10:42And they have a peer-to-peer exchange. They have a spot exchange. They have a debit card that you can use for spending your coins. And they also have a gift card marketplace where basically if you're signing up to NoOnes and you're selling stuff on there, you're giving these people more liquidity and more stuff that they can basically sell to one another to bypass their fiat currencies and try to live in a way that grants them more financial freedom. So check it out. It's n-o-o-n-e-s dot com.
2:11:24And ironically, it does have on the main page some graphics that look like they were made on the NES. Maybe there's a joke here and I don't get it, but it's called NoOnes, not No-NES dot com. Check it out. And now I still have some questions regarding the design of the Quai Network. For example, you mentioned something about the supply of the Quai token, which is deflationary. Can you explain how that works from an economic point of view?
2:12:00Karl KrederYeah, so the Quai token has a deflationary but dynamic issuance. So Quai is emitted proportional to the log of diff. So as the difficulty increases, the amount of Quai emitted per block or sorry, the amount of Quai per hash continues to go down over time. So with the growth of the network, the amount of Quai emitted asymptotically approaches a value. So it's not a cap ceiling, but the inflation rate tends to zero with time.
2:12:51Vlad CosteaOkay, so basically there is no cap.
2:12:56Karl KrederYeah, so the reason that we didn't want to do just a strict cap to value is, you know, it was great what Satoshi did with four-year halvings. But the day that we live in, we can't adequately predict the rate of growth of a system. So the time constrained emissions decrease is inefficient relative to adoption, would be my statement. So it kind of creates cycles and within those cycles, the sort of shock in the system makes it so that there's periods of time where it's like more favorable to buy Bitcoin and less favorable to buy Bitcoin.
2:13:54And so when I'm coming into the system as a new user, you know, and typically when new users come in is when it's probably less favorable to buy Bitcoin. So in some ways it creates interest in sort of that run-up, but maybe it's not as efficient as it could be in converting those people into long-term holders. And in terms of capital deployment, you know, people are saying, I might want to sell my Bitcoin because there's a cycle and I might want to buy until I think the cycles hit a bottom.
2:14:31By making the emissions responsive to the hashrate, it should always create a deflationary environment relative to growth. So then the statement would be any time should be a good time to buy if you think the system is going to grow.
2:14:49Vlad CosteaOkay, but to a lot of people, basically the reason why they got into Bitcoin in the first place is this 21 million meme, right? Because they know, or at least they assume that there will be no changes in the future to the supply of Bitcoins. And is there any math behind this? Like, do you believe there's going to be like this amount of Quai tokens in the upcoming future? Or is it so dynamic that you can't fully predict that it goes up and down according to?
2:15:27Karl KrederWe can kind of bound it. I mean, it's all predicted based off like the estimations of growth rate of the hashrate. But the estimations that we make put it somewhere between 12 and 17 billion.
2:15:41Vlad CosteaWow. Okay. And did you aim for this number or is it just basically what happened?
2:15:53Karl KrederSo, I mean, on a very gross basis, we aimed for a number in the billions because we think sort of one thing that maybe isn't advantageous to Bitcoin is the cost per token. I mean, obviously it's divisible, I get that, but there's some number of buyers that get put off by Bitcoin because of that. And it creates more difficult pricing for any sort of transactional use because we're always dealing like a high degree of decimals where humans aren't really conditioned to operate past three decimals. And they're also not really conditioned to operate sort of like in hundreds of thousands of sats.
2:16:38So it's kind of in this like a numeraire, no man's land where it's like cognitively stressful for one reason or another for humans to deal with. So we think like the less cognitively stressful bound is having billions of tokens rather than millions of tokens, I suppose is the statement. So we certainly made that choice in the system. But in terms of the outcome of the total number being, you know, 12 billion or 17 billion, that's kind of just the outcome of the math and what we think the curves might look like.
2:17:15Vlad CosteaThere was someone in the chat who goes by the name of crev23 and he made a very bold statement here. He said, make me want to use your shit and it will be the greatest coin ever. It doesn't mean coerce me. That means make me want to be part of it, to contribute to it, incentivize me. That's how Bitcoin got where it is. Hashtag game of incentives.
2:17:40Karl KrederNo, I agree with that statement, 100 percent. So, I mean, that's that's kind of where we focus some of our efforts now, because for the most part, we're not really doing protocol development. The concept was to try to get the protocol done. So there's a fork coming in that's going to like add a couple opcodes to the EVM. But the most of our time is actually focused on business development, like marketing and product development. You want you want to do maybe a live demo?
2:18:17Vlad CosteaYou're going to do the live demo, right?
2:18:20Karl KrederWell, I'm going to try to get you to do it. You have an iPhone or an Android?
2:18:24Vlad CosteaI have an iPhone right now, and it keeps ringing because it connects to my laptop to become the default webcam and I'm trying to prevent it from it. But it keeps reminding me that it can.
2:18:37Karl KrederWell, if you're interested, there's a mobile wallet and it's called Blip Pay in the app store. And if you download it, we can do a live demo of how fast the distributed network works.
2:18:55Vlad CosteaBlip Pay?
2:18:56Karl KrederBlip space pay. Yep. B-L-I-P space P-A-Y.
2:19:04Vlad CosteaOK, so it's this one with the green logo.
2:19:09Karl KrederYes, sir.
2:19:11Vlad CosteaI'll try to get it.
2:19:15Karl KrederSo if you get it, don't worry about referral code or importing. Just create a new wallet, don't put a referral code in and the display name only exists locally. So in this app, we don't like have a back end. We designed this so that the only back end in this app is the blockchain. So if you get it loaded, you'll be able to show me your QR code. I'll be able to send you something and within that, like five to eight seconds, you'll be like, aha, I'm now a proud owner of some Quai.
2:19:49Vlad CosteaSo there is a terms and conditions, which I'm going to pretend that I have read. And there's also a referral program, which I skipped. Skip, skip for now. So setting up wallet, what is it doing right now? Is it downloading from the blockchain?
2:20:06Karl KrederNo, it's just it's deriving an HD wallet for you and a seed phrase. And then your display name is just whatever you want it to be, doesn't, doesn't go public or is it just something it only exists in the QR on the phone. So when I scan your QR, it's gonna say like, I'm sending money to whatever that name is. But it doesn't go on the blockchain, it doesn't go in a database, it just gets encoded in your request QR.
2:20:37Vlad CosteaOh, this is interesting, but horrible for privacy.
2:20:41Karl KrederYeah, well, you can make it anything you want. So you can make it anything you want, right? Like, it's just for if I'm going to interact with someone, and you're gonna be like, Oh, send me some money scan my QR code, it'll come up and say, Vlad the destroyer, or like, whatever, whatever you want to put.
2:21:00Vlad CosteaOkay. So it seems like I'm set up.
2:21:04Karl KrederOkay, so go to the pay, the pay button in the middle there. And swipe to receive, give me your QR code.
2:21:15Vlad CosteaAm I getting bribed right now?
2:21:18Karl KrederLet's see if we can get it on your camera.
2:21:19Vlad CosteaI got paid for this interview.
2:21:26Karl KrederBack a little bit more, a little bit more. Back a little bit more. Oh, no, I don't know if I'm gonna be able to get that. It's not. If you go too far forward, it's
2:21:26Vlad CosteaWait, let me focus the camera.
2:21:26Karl KrederOkay. Yeah, if you focus on it, like we'll be good. Oh, yeah, right. Right there. That should work. A little more focus. We can do it.
2:21:54Vlad CosteaOh, man. I'll screw you iPhone. So I can't really tell if this isn't focused or not.
2:22:04Karl KrederOh, go forward a little bit. Right. Back, back, back. Right there. Come on. Okay, maybe just drop the address. Unfortunately, we're not going to work. Wait. Oh, okay. Go left a little bit. You're off screen. Oh, right there, right there. Come on. Oh, there it is. Boom. We did it. We did it. And you'll have to refocus on your face now. Once you do that, I will show you. So right now I have you and your chosen name, which is Vlad. I'm going to send Quai.
2:22:40Vlad CosteaIs this like a BIP-47 PayNym?
2:22:44Karl KrederSo Quai is BIP-44. Qi is BIP-47. So Qi is the privacy one and Qi is default PayNym. So I'm going to send you 100 Quai. I'm going to hit confirm. And so if we look, it's sent on the network. And so if you go to your wallet tab and hit refresh, you should have it.
2:23:16Vlad CosteaNot yet.
2:23:18Karl KrederRefresh one more time.
2:23:22Vlad CosteaStill nothing.
2:23:24Karl KrederMaybe click off wallet and back on wallet. It should show.
2:23:31Vlad CosteaSo I'll go to contacts and now wallet and refresh. Still nothing. No balance. We should count the seconds.
2:23:43Karl KrederThis is longer than eight seconds right now.
2:23:47Vlad CosteaWell, there's more to it than...
2:23:49Karl KrederSlower in Romania.
2:23:51Vlad CosteaBlockchain, yeah. You have to deal with Wi-Fi and VPN and everything else.
2:23:57Karl KrederOh yeah, that too.
2:23:59Vlad CosteaLet me close this and open it once again.
2:24:02Karl KrederI see it on the explorer already, but maybe you're not getting it on the phone.
2:24:09Vlad CosteaI closed the app. I've opened it once again. Still it says I have no transactions.
2:24:16Karl KrederIt is here, which is on the blockchain. Is there like a weird setting you have that's not letting you get connectivity to that app?
2:24:33Vlad CosteaNo, I don't think so. Interesting. It's always funny with these live demos, right? Because you never know what's going to happen. But it's still zero. Still zero?
2:24:53Karl KrederInteresting.
2:24:53Vlad CosteaI know close to my face, but it's too small.
2:24:56Karl KrederYeah, yeah, yeah. Okay, interesting. So I sent it. Just out of curiosity, what is your address? Maybe it didn't encode correctly.
2:25:08Vlad CosteaIt's 0x003d and ends with B19a5f.
2:25:18Karl KrederOh, that's interesting, sir.
2:25:22Vlad CosteaSo did it scan the correct address?
2:25:22Karl KrederI don't think it scanned the correct address.
2:25:22Vlad CosteaSo where did you send the coins?
2:25:30Karl KrederApparently into the ether. That's wild. It has checksumming. I've never seen that before. But I've never also tried to do that QR code before.
2:25:42Vlad CosteaSo what are the odds of this happening? To get an address that gets the checksum?
2:25:48Karl KrederI have no idea. That is... I'm a little bit floored by that. Is there a way you can send just the address in the chat or something?
2:26:03Vlad CosteaYeah, I can try that.
2:26:06Karl KrederOr if you want to just copy your QR code and send it to me by email, I can do it that way too. Just this little camera show we have, I don't think it's working.
2:26:20Vlad CosteaThis is funny. I can just send the address to a Twitter DM.
2:26:25Karl KrederOh yeah, you can do it that way too. Yeah, so the communications... Notice how we're using other communication platforms, right? There's no backend. We don't know who the users are. We can't help you connect the users. Your only ability to connect to each other is through your existing networks and QR codes.
2:26:48Vlad CosteaSo did we just burn 100 Quai from the 60 billion supply?
2:26:54Karl KrederYep, deflationary right there. There we go. Okay, let's see if I can get this address now. Copy, baby. Copy message.
2:27:09Vlad CosteaJust let me know when you send it so I can show it to the camera that it's not zero anymore.
2:27:15Karl KrederYeah, okay. So here we'll do it again. I'm going to confirm here. Oh, there it is. So that says successful. Should be able to view this guy. Oh, maybe it's my end. Hold on, what is going on here? Oh, that might be the problem.
2:27:53Vlad CosteaFYI, still zero.
2:27:55Karl KrederNo, no, I know. I know. I think this might be on my side. It's not actually making it to the Explorer for some reason. I haven't had that happen before.
2:28:18Vlad CosteaDo I get a bug bounty or something?
2:28:21Karl KrederYeah, man, I don't know. See here. I think it's because I had my VPN on. It wasn't actually going through. I have some special software on here. I might have to restart this. Anyway, let's keep talking. I will restart my phone and we will try this again.
2:28:53Vlad CosteaThere's a Kaspa guy in the chat who's really making fun of us right now.
2:28:58Karl KrederOK, OK. No, it's my phone. My phone. It does magic shit. I'm going to do it just from my other thing. All right, so I'm going to do this via Pelagus, which doesn't have the same issues that my phone has because I'm a paranoid person. And we will. You sent that in the Twitter DM. I will grab that from the Twitter DM. I will tell you when I send it.
2:29:42Vlad CosteaThere's someone who said, don't worry, bro. It even happens with the lightning network, right?
2:29:49Karl KrederNo, it's not the lightning network. It's my phone, but all. OK, so I'm pasting the address right now and I'm hitting send. OK, I have sent the transaction and it has actually gone over Wi-Fi. Unlike my phone. And it is pending on the block explorer.
2:30:07Vlad CosteaOh, I got it. OK.
2:30:08Karl KrederAnd you got it. Right. So that's. Yeah, I won't go through what my OPSEC is, but that was an OPSEC problem on my phone. So when I don't have those problems on my phone, that's how quickly it goes through.
2:30:27Vlad CosteaSo explain to me what happened in the background. Like, let's look at a block explorer and see how this 100 Quai transaction actually took place, like from which shard and when this is going to become final, right? Because I know you said you have the prime chain, which is the one at the top. And I guess a transaction is not considered final until it's written in the prime chain.
2:30:52Karl KrederYeah, so no, no, no. A transaction is final right now, actually. So this this is the transaction.
2:31:07Vlad CosteaOh, I guess he left by accident.
2:31:11Karl KrederSorry, guys. I hit I hit refresh on the wrong window. I thought it was on the block explorer. Yeah, sorry. Try to be. Yeah, here we go. So, yeah, you see that. So that is already 33 blocks deep. So that is like a final transaction. And that also has already had prime blocks put on top of it. So in one shard setting, you see this five to eight second topology, and you'll always see that five to eight seconds with an intrashard system. But even for inter-shard transactions, you'll still get five to eight second confirmations.
2:31:59So you can trust that you'll get that money. You just won't be able to spend it depending on how many shards are in the system for that inter-shard time. So that's why we say the difference between, you know, a confirmation and settlement, meaning you're able to respend the funds. So even if, again, you were in this 100 shard setting, you would still get that five to eight second confirmation. It would just be you wouldn't be able to respend that for potentially 90 minutes. But in that 100 shard setting, that means you're doing over 100,000, maybe a quarter million TPS.
2:32:33Vlad CosteaOK, so it comes from an address that starts with 001. I guess that's on the first shard. And my address is on the third shard.
2:32:44Karl KrederOh, no. So currently in mainnet, there's only one shard still because we haven't saturated the shard. When the protocol detects saturation via the uncle rate, it automatically instantiates a second shard. So you go from one shard to being mineable to two shards being mineable. And then once those saturate, you'll go to four shards. And then once that saturates, you'll go to six. And then it just keeps automatically increasing the number of shards with sustained demand.
2:33:20Vlad CosteaYeah, so someone just sent me 8 Quai, apparently. And I see that it also has like a price function. So this is like four dollars that you sent me. So anyone who's going to say I got paid for this interview.
2:33:40Karl KrederWe got paid four dollars. It's going to be infinite in the future.
2:33:47Vlad CosteaI'm going to go buy myself ice cream after this interview. Yeah, and someone sent me like thirty four cents, which is 8 Quai. I guess someone else scanned the QR while we were struggling.
2:34:03Karl KrederThey managed to scan it and send it, unlike me. Let me see if this reset. Mine can try it here again.
2:34:13Vlad CosteaNo, it's all good. You don't have to send me any more of this. But I'm still curious. What is the way to acquire Quai, right? Are you guys listed on any exchange? Because maybe it's easy to get your hands on some coins who are going to start mining or something, but there needs to be a market for it.
2:34:30I fixed it.
2:34:30Karl KrederI fixed it. You just got 10 more Quai.
2:34:39Vlad CosteaNot yet.
2:34:41Karl KrederRefresh one more time.
2:34:44Vlad CosteaI did.
2:34:45Karl KrederYeah, see, OK, that was that was that was my issues over here in terms of getting getting a broadcast. But cheers. That's proof of work. Decentralized proof of work at scale. And it will keep doing that if everybody on the planet decided to use it.
2:35:11Vlad CosteaI'm reading the comments in the chat. They said bros dying inside. Poor guy. When you couldn't get the transaction proof.
2:35:24Karl KrederWell, yeah, I mean, it wasn't the blockchain. I again, I'm not going to, like, reveal what my OPSEC is, but I have OPSEC. It was an OPSEC problem on my phone. I have certain things that I do in certain ways. That's all I'll say about that.
2:35:42Vlad CosteaOK, is it possible to create a BTC token on the network and be able to basically lock Bitcoin on the Bitcoin network and then issue it on the Quai Network and be able to send it this fast with your network and paying the gas fee with your gas token?
2:36:05Karl KrederSo Portal to Bitcoin is integrating. They haven't completed that integration yet, but when they do that, that will be possible.
2:36:15Vlad CosteaIsn't Portal to Bitcoin doing some sort of L2, which is very different?
2:36:23Karl KrederSo the purpose of Portal to Bitcoin is to bridge Bitcoin to other more composable networks. And that mechanism, if you want to call it a bridge or an L2, right? So if it's like a federated bridge, OK, it's a federated bridge. If it's a bridge that has sort of like a consensus mechanism, then you're kind of getting into sort of like L2 territories. But their purpose is to bridge to other networks. So bridge Bitcoin to Solana, bridge Bitcoin to Ethereum, bridge Bitcoin to Quai. And so once that integration is complete, you'll be able to bridge Bitcoin over to Quai.
2:37:12And you'll be able to send that Bitcoin as fast as we just did that Quai transaction.
2:37:19Vlad CosteaRight. But the question is, where can you use Quai today? Or are there any plans to integrate this to add this on exchanges or swapping services like SideShift?
2:37:30Karl KrederYeah, so it's currently traded on MEXC, Gate, and like 10 other small, smaller exchanges. We're also on the Kraken listing roadmap. And so we're in the process of getting integrated with Kraken and then U.S. And European customers will have access in USD and Euro pairs. And then there's another thing that's happening with a major bridge. And we're hoping we get to announce that next week, which will also enable you to bridge to Quai pretty much from anywhere.
2:38:24Vlad CosteaSo what else should I ask you about this? Why should anyone pay with Quai? Like, what's the pitch?
2:38:36Karl KrederYeah, so I guess the point is that it's proof of work. It's decentralized and scalable with scalability. By maintaining EVM programmability, you can build any sort of application you want. And as that application grows or that use case grows of its payments, the chain will continue to increase throughput using the sharding, guaranteeing the users and the developers fixed low-cost fees in perpetuity. So if we look at CryptoKitties from 2017, it caused gas prices to spike and Ethereum to grind to a halt for a simple NFT cat breeding app.
2:39:26And one thing that sort of constrained how far and how long they grew was the fact that people couldn't use the network anymore. There's been other examples of this. If you look at payments in places that have hyperinflating nation state currencies, like Argentina, a lot of people in Buenos Aires were trying to use BSC for payments in lieu of the Argentine peso. But as they started to do that, transaction fees went up and it priced them out of market. And then they said, okay, we'll move over to Tron.
2:40:05And then when they started to do that, transaction fees went up and priced them out of the market. The idea here with Quai is that because you have scale, if you get a major city or country or geography to adopt payments, when they do that, they'll continue to have the same low fees and the same fast processing time. So yeah.
2:40:30Vlad CosteaThe other question is, how do you guys get paid for working on this? Like, is there any dev fund or is there any pre-mine or are you funding this out of your own Bitcoin stashes, which you're actively selling to be able to sustain your work?
2:40:47Karl KrederSo we took venture capital, right? This was a three year effort with a team of about 10 engineers and then an equivalent number of sort of marketing and business development guys to get this to fruition. But yeah, I'm personally invested. I am also funding aspects of it with stashes of Bitcoin. The investors were Polychain was the lead, as well as Alumni Ventures, a couple other smaller ones, but so there's the pre-mine to cover that, a team allocation. There's also a foundation allocation, which is used for development, but as well as listings.
2:41:30So I should say adoption incentives. Part of that is going to be used because the world that we live in isn't fair. So if you look at something like Kaspa, Kaspa is having tremendous trouble, say, getting listed on certain major exchanges because they don't have a fund to deal with. We have that fund. So we expect that to happen before it happens for Kaspa. And that's kind of what we allocated that for. So the whole point of the foundation funds is to realize that we live in an unfair world and effectively bribe, I shouldn't say bribe, incentivize existing power structures, user bases and companies to adopt Quai with those funds and to do so in such a way that you're always bringing in more aggregate value to the chain than you're paying.
2:42:32But we see that basket of funds as being important for getting adoption into the existing power centers.
2:42:44Vlad CosteaMakes sense. But where I was basically hinting at was a question about the launch of the Quai Network. Did you guys take out any amount of coins before launch or did you benefit from the information asymmetry, meaning that you are the first miners and therefore you were able to accrue a certain amount of coins before the rest of the world joined in?
2:43:10Karl KrederOh, I mean, this was pretty well publicly announced. We almost had a day one Binance listing, but we kind of got front run by another project. When we launched, we had over 300,000 followers from Twitter. We'd been sort of actively advertising this for two years. And when we launched, the hash rate was as high as it's ever been. The hash rate was three terahash in the first day. So I think it's as fair as we could possibly have done it. But if you kind of look at the emissions, our emissions curve didn't highly advantage day one miners, I put it that way.
2:44:03And actually, if you mine today, you're going to be doing better per hash than you were if you were mining at launch because the hash rate was so high immediately after launch.
2:44:14Vlad CosteaAnd something else that happens with new coins is that at launch there is hype about the price and everything and the supply is low and therefore it's very easy to push the price to a certain level. And later on, if the supply gets higher, even if there are more users and there is more economic activity, it becomes more difficult to reach that price. And people will say, well, it was a pump and dump. But what really happened is that the supply that was available in the first few days was very low.
2:44:50And there was some hype fueled by the marketing and the announcements that the network or the coin is launching. And it leads to the situation like I'm pretty sure that, for example, Monero right now is better off than it was six or seven years ago when it reached its all time high. It's just that the supply was lower at the time and there were more coins that were mined and it's difficult to push the price to that point.
2:45:21Karl KrederYeah, I mean, if you kind of look at proof of work launches, they have sort of identifiable traits, right? If everyone knows about the launch, you kind of get like an instantaneous spike, then you get sort of a consolidation period before maybe you get market discovery of price. If they don't know about the launch, you don't get a spike in price. So that lack of a spike in price at launch is indicative of people not knowing about the system when it started getting mined. You know, and then then maybe the consolidation period is kind of just like flattened to the right because no one really speculates on it initially.
2:46:06And then at some point people eventually discover it and you get price moving going up. But I would actually say kind of the price spike that we saw the week after we launched was indicative of people knowing about the network, especially considering the timing of it. Like the timing of like our launch was absolutely terrible in terms of like the market. Yeah, on like a number of fronts, but such is life.
2:46:36Vlad CosteaWell, what was the reception like? Did you get any reviews or opinions from people in the industry who may or may not be famous?
2:46:49Karl KrederOf like the launch or the technology or what?
2:46:49Vlad CosteaI guess both.
2:46:49Karl KrederYeah, so there's not many proof-of-work people that exist that are active and notable. The majority of people are proof of stake people, but within the family of proof of work people, we've obviously done work ourselves. We've also worked with people like Dionysis Zindros. We worked on programmable proof of work as well as non-interactive... Or sorry, proof of proof of work and non interactive proof of proof of works. And we use that heavily within Quai for a number of different reasons.
2:47:46But, you know, we did a security proof with them. We've published like six peer reviewed papers with probably 20 different contributing co-authors, a lot of whom are professors at universities. And then even within the Kaspa ecosystem, we work with Shai and we're currently going back and forth on maybe what I'm considering to be like a grand unified theory of proof of work in some ways. But yeah, maybe I can just briefly touch on that one, which is maybe a thing. I've circulated the paper in kind of like a research channel if people are interested in finding it, but it's still in draft and we're working on it.
2:48:38But basically when it comes to proof of work, all you're trying to do with consensus is sample. So the hashing is a physical process and consensus is trying to understand where 51% of the hash is voting effectively. So you have the parameter, which is the hash rate, and based off your sampling, you're trying to measure the hash rate, specifically where the hash rate is pointing. And so if we look at what that means is that to know that 51% of the hash rate is voting on something, you have to have enough samples to say that that indeed represents the population.
2:49:24So if, for example, I was trying to measure like the height of males in America and I took six samples, if I look at the outcome of that sample relative to the true mean of the population, in a sample of six, right, I might come up with the answer that the average population height is five foot two. I might come up with the answer that the average population height is six foot four because my sample size is so small. So I don't actually know. Sort of what the measured parameter is.
2:50:00There's a lot of variance in it. And this is the same with hashrate. So in Bitcoin, we kind of say that things are finalized after six blocks, which is actually really saying that the confidence interval that six blocks represents the majority of the hashrate is actually only 87%. And with Bitcoin, you know, because the times are so long, maybe that doesn't matter as much, but when you compress times down to trying to do five to eight second transactions on five second blocks are potentially even faster than that with workshares, you want to sample the system much more quickly.
2:50:40And if I sample the system much more quickly, I am able to ascertain where the majority of hashrate is pointing much more quickly, which then allows me to finalize the transaction much more quickly, at least from a statistical basis. So the process of proof of work blockchains is really just a process of sampling and you want to sample based off the available network resources and do so in the most efficient way possible. So basically the idea of this paper that we're working on is looking at the information theoretical process of sampling using something called Fisher information and then understanding how quickly you can ascertain the process parameter within sort of a certain window of variance, i.e. the hashrate, most efficiently based off what information is coming relative to the sample population. So this actually creates implications on ideal graphs as well as ideal consensus measurement algorithms. So one conclusion that I'm getting to is that the best graph type in the system is something that we would say is maximally anti-chain. Another way to put that is we want as few dependent events possible in the graph to maximize the information that we're sampling. So I can actually, using sort of this information theoretic treatment, look at two graph types of proposed blockchain systems and say one is demonstrably better or worse than the other in terms of achieving consensus.
2:52:30And what I'm really trying to get to is to convince Shai of this, and if I convince Shai of this, I think it'll be very reputable to say that a structured DAG using workshares like we do in Quai is better than an unstructured DAG that they use in Kaspa.
2:52:55Vlad CosteaI feel like I just opened the floodgates for some sort of flame wars between the two communities, but there is someone in the chat called Francisco Savia who says Vlad asked about kipper money to DRK. I don't even know what this means, but you might know.
2:53:14Karl KrederYeah, so one of one of the things right you mentioned earlier is that we've built the system and you can see that it can function very quickly. And we've demonstrated it publicly at scale. Maybe you can believe us or you can verify it, whatever you want to do, but you know, that's interesting, but the point that you were raising is in 2025 maybe people have moved past the idea of money and you're just thinking about speculative assets. So what we're thinking about in our ecosystem development as well as our internally developed or pushed products is serving underserved markets in a way that we sort of hide the crypto as much as we possibly can, if that makes sense.
2:54:07Because what we're saying is we have a digital form of money that can be used in ways that sort of previous predecessors weren't able to be used. So one sort of group that we actually see as marginalized or disenfranchised is content creators because the content creators are owned by the platforms. They have a very high threshold to monetization. Some types of content are disallowed to be monetized, one of those being crypto depending on the platform and basically what we're saying is we've created a tool which we've first integrated into X, which allows you to use X, but it adds the feature of being able to tip in X, if that makes sense.
2:54:58So you just sort of use the Kipper app with the Kipper extension either on desktop or soon to be a release of mobile and it looks just like Twitter to you, but there's another button and when you press that button you can tip a creator in Quai, in SOL, in USDC, in Beam tokens. And that's sort of Kipper, but the bigger view of Kipper is ideally making it so it's multi-platform and what should happen then is you can actually make it so that the the following and the followers become part of the blockchain, the monetization becomes part of the blockchain and then what a platform means to a creator is simply a data availability layer.
2:55:44Ideally, if you can take the money out of the platforms, the creators aren't dependent on the platforms, they're dependent on the audience that they can cultivate through the blockchain, but you're bootstrapping it with the audience that they have on the platform, if that makes sense. So if we look at something like Farcaster or Nostr, I think it's very difficult to try to bootstrap a broadly accepted secondary social network. What we're saying is let us use YouTube, let us use Twitter, let us use Farcaster and you can use it all in the same app with a unified monetary layer which you don't need to sort of build a new community, a new network effect for.
2:56:25That's what Kipper money is. So right now, that's like the big vision, right now very simply it's just tipping in X and if, give me a second, I can probably gin up a little demo here.
2:56:38Vlad CosteaIt's like the good old tipper bot that people were using on Reddit and I think also Twitter for a while. And at one point there was something called BottlePay that they made on Bitcoin using the Lightning Network. It was fully custodial, but you could integrate basically with a plugin to send Bitcoin on the Lightning Network to a bunch of YouTubers or whoever was willing to install that plugin. And then of course it got shut down by regulators.
2:57:13Karl KrederDid it really? I'm actually not familiar with that. So, I mean the interesting thing here is that it shouldn't be sort of able to be shut down by regulators because everything you know is happening on chain. It'd be interesting to hear more. What was the one that was shut down by regulators called?
2:57:39Vlad CosteaBottlePay. They were based in the UK and that's where they raised their VC funding and I think that that was ultimately their downfall. Because they had to deal with more strict KYC AML requirements, which they could not. So they decided it's better to shut down the service. But with it, it was some of the most fun I've ever had on Twitter?
2:58:07Karl KrederYeah, I mean it is kind of addictive on Kipper. Let me just Let me just see if I can I can tune up a demo real quick. One second. I've been testing so like We've been testing like a release. So I just got to make sure we have like a working extension right now. Let's see.
2:58:46Vlad CosteaThere is someone in the chat who told me to put my Quai address in there. I mean, I'm not gonna say no to free money.
2:58:59Karl KrederOh, they're gonna tip you, huh? Let me see if I can get this real quick. One second. Actually, um, let me open a different profile that I know is working. Just share that one. I will.
2:59:36Vlad CosteaI guess this is how you know that this is not scripted, right? We had no plan whatsoever.
2:59:42Karl KrederYeah, I mean, I mean we like agreed to this. I don't even know like a day ago. So it's all good. Okay, so this is kind of like a throwaway account for me, but this profile I was testing with me. One second. I'm just going to set up real quick. Next. It's all this guy. Yeah, I mean, unfortunately, like I literally was testing this on like this morning. So it's just a little bit of a state. But just give me one second.
3:01:17Sweet. Sweet. Okay, this should work. There's X, there's X. Let's just check it. We have a thing. We have a tip. Okay, sweet. Cool. Let me see if I can share this screen now. We're getting here in this one. I have a different profile. One second. Okay, this is the profile. I'm gonna have to share a screen. Another profile, which is this one, I believe.
3:02:06Is this? It's just this one. Can you see me moving a cursor here?
3:02:18Vlad CosteaYes, I can.
3:02:20Karl KrederSweet. Cool. Okay, so this is Kipper. And what you see here is actually, if you notice these top three tweets, they're, they're like highlighted in green a little bit. Those are promoted. So based off like who's sort of the most tipped at the moment, they'll show up here at the top. But anyone, you know, in X, you can go to. And, you know, like I said, this is just like a test account. So I don't have like a bunch of good follows here. But I can say I want to tip with Quai, I can put five Quai in, I can send a tip.
3:02:58And then this just automatically tips out of what's called a Kipper jar. So it's kind of like a permission. It's a wallet that you're giving permission to the app to use. And so the reason it's called a Kipper jar, because this isn't meant to be like a highly secure wallet that you put a bunch of money in. You're just going to put money in for this, this concept of tipping. And with that, then you don't need to sort of confirm transactions sort of every time we make them.
3:03:27So like if I go to say, you know, let's, let's go to you, Vlad.
3:03:34Vlad CosteaIt seems like I'm not the first option that's trending right now.
3:03:43Karl KrederNo, no, it's the what's what's your full handle? All right. This is like I said, this is a throwaway account that like I created.
3:03:51Vlad CosteaI think that's an easier way to the Vlad.
3:03:58Karl KrederYeah. There we go. Cool. Right. So here's, here's you. Here's our space. I'm going to just go ahead and tip us. I'm going to make us trend. I don't know how much Quai I have. Let's do 10. I can't do like 100. I don't know if I have 100 in the test account. But so like, I've tipped you. And now that isn't a smart contract. And when you go and sign up for Kipper, you'll associate a receive address with your handle, and you'll be able to pull the money out of that contract.
3:04:35So that's like an escrow.
3:04:37Vlad CosteaBut if we go to someone that we know is like signed up, like, let's say myself, well, you did send some coins to Elon, and I'm not sure if he's gonna get on board. So what happens if he doesn't claim his deflationary? It's deflationary at that point. I was thinking maybe that there is a time lock and after a certain amount of time, if he doesn't claim it, it goes back to the sender.
3:05:04Karl KrederUm, yeah, I mean, that could potentially be an upgrade. So for example, if I if I go to like myself, and I send a tip, it just goes through like that. And then that has like already happened is in my like, receive wallet that I have set up for Kipper. So, but it's pretty cool. Because like, when you're on X, and you can actually also set it up. So you can like tip with likes. So I'll see if you can do that. Just change this, this like release, actually, I think I have to go here to do it.
3:05:44But you can you can set it up to also do likes, like, I can set that up so that for like, I can set a default amount. So every time I like I could like tip you like 100th of a Quai or one Quai or like whatever it is. And then you can kind of see like the feed right of the tips that like just went through. And then tips that were happening before that, sort of in the system, and then you can see some stats on it.
3:06:20But this is kind of like the, yeah, an application that we're pushing, we're going to be actively pushing, we've kind of been developing it to get it to the point where it's like a truly releaseable marketable product. Where it's not just Quai, but it's also like Quai and Solana. And then we're also looking to go multi platform, and mobile, which, you know, mobile, hopefully, we can like TestFlight it out next week. And then we're gonna start adding other platforms, we're probably going to do Instagram first. So one of the other sort of features that we see that's like interesting with I gotta find my okay.
3:07:09Okay, we're back over here. So one of the other features that's interesting with the sorry, I'm distracted because I was like trying to do the demo. I don't know what point I was trying to make anymore.
3:07:27Vlad CosteaSomething about features.
3:07:29Karl KrederOh, yeah, no. So one thing that's like cool about Kipper for me is that it's Web3, but it feels like Web2. And it's happening on the main chain. So it's not like we're doing this in a, you know, state channel, or, you know, something of the such that's just on chain activity. And it can happen at the speed that you kind of use a Web2 application, and you don't need to sort of do anything special to get it to work. So yeah, I think it's kind of fun, just as an example of applications where you're trying to sort of solve a problem.
3:08:05And I guess the approach we're taking is it's not about you know, trying to rebuild social networks, it's about allowing people that have activity on those networks to be able to maybe more directly monetize them.
3:08:27Vlad CosteaRight, so someone suggested in the chat that I put my QR code on the screen for people who want to send Quai. I don't have any objections to that you can basically use this, which I just put on the screen. And I'm going to announce whenever someone sends me, I got like 110 Quai so far from people.
3:08:27Karl KrederWe, we need to add a ding.
3:08:57We should like ding this guy, you know, we actually, it would be actually cool if you could message within the transaction.
3:09:05Vlad CosteaLike, can you actually do this to send a message to the person using the transaction?
3:09:11Karl KrederUm, not with the transaction, but there's a group that's building an application on Quai, it's called Talking, which it uses transactions to send messages. So one cool piece of Qi is Qi has fixed denominations. And it also has trimming, or sort of dust-collector pruning, I should say. So when you use Qi, you know, this is the token that is UTXO based, it's fixed denominations. It's meant to give you like a modicum of privacy, right? Cash like privacy, specifically. One second. Getting a little sunny in here.
3:10:11It's meant to have cash like privacy, but because we have fixed denominations, and those fixed denominations go down to what we call a Qit. A Qit is effectively, maybe you should think of it as like a thousandth of a dollar, roughly speaking, based off like Qi prices and everything. So that's kind of like the minimum denominating unit of the Qi ledger. So it's the minimum denominating unit. It's the equivalent of a SAT, I guess. And then the ledger goes up in ones and fives, and I think there's 17 denominations.
3:10:42But it was designed to basically each denomination becomes an anonymity set when you're using them. But the issue is the denominations that are under one Qi could bloat the ledger just in a huge way. So there's the issue that if you have like a lot of Qits lying around, we don't want those to like lie around indefinitely. So anything that's under one Qi has only a temporary use time in the ledger. And after, you know, a denominated, or sorry, a denomination determined duration, that denomination, if you don't spend it is trimmed out of the set, if that makes sense.
3:11:28So I kind of think of this, like, if you're old enough to have change in your life, and you would take your change and kind of put it in a change jar. And then, you know, every three months or something, you would take it to the bank, or you take it to in the US, we call it like Coinstar. And you take all your change, and you get, you know, some money or Starbucks gift certificate for it or something. We kind of think of sub-Qi values as the same idea, meaning, if you want to do a business, and that's valuable to you, that's fine.
3:12:01Just collect and re aggregate your small denominations prior to expiry. If they're not important to you, your change jar kind of eventually vanishes from the ledger. But that keeps the ledger clean enough that you can use it transactionally by still maintaining privacy and get you know, adequately precise denominations. Now, the consequence of that, though, is that Qits are able to have a little bit of data with them about 20 bytes. So if you publish a Qit, with an encrypted message, you can chain a bunch of these things together.
3:12:38And they only have a residence time in the blockchain of like three weeks. So it effectively creates a perfectly private routing system for how you can transmit messages to another party, which is kind of cool. So even if you use Signal, you know, which is using the Whisper protocol, which is peer-to-peer encrypted and double ratcheted, you're still suffering from the fact that my IP is sending information to your IP. But with Qits, it's basically you're you're publishing them to you can think of it like as a message board.
3:13:14And everyone that's running a node is kind of walking by that message board. And you can think of them as like, you know, encrypted notes on sticky pads, but the sticky pads fall off the message board every three weeks. So everyone can like walk by the message board, nobody knows who's talking to whom we never associate our IP addresses, because it's mixed in not only with all chain traffic, but all the messaging traffic. And you could actually potentially have something that's better than the Whisper protocol, because you can fully obfuscate peer-to-peer, or sorry, IP matching.
3:13:50Anyway, it's kind of cool. It's called Talking. So it basically is like using that property to do messaging using transactions on the chain.
3:13:59Vlad CosteaI know that Zcash has something like that.
3:14:05Karl KrederIP obfuscation, or just messaging by transactions, messaging by transactions, which is pretty private.
3:14:14Vlad CosteaAnd also, I guess there is a character limit to it. But it's pretty cool anyway. And the benefit of it is that they can have wallets in the app stores, which are labeled as just messaging apps.
3:14:36Karl KrederOkay, it's like, it's like a backdoor app store thing. That's funny. It's, it's unfortunate we didn't get more adoption of PWAs in this world. Because that also is an option for people. Progressive web apps. Yeah. Base had something initially was like the base wallet. It was a Coinbase wallet that was a PWA, but most PWAs don't get traction just because it's not in the normal workflow. But it is completely independent of an app store, which is nice. We've thought about releasing Blip as a PWA too. We just haven't had the resources and time and priority.
3:15:26Vlad CosteaI mean, the advantage is that it's more compatible with pretty much every device. You can even open that on a BlackBerry, I guess.
3:15:35Karl KrederYeah, it's compatible with every device, but it's independent of the app store. Exactly. So if you can manifest it as a PWA and they're like, Oh, we don't like you for whatever reason, doesn't matter. No one is controlling it. It's just, if you can go to a web address, you can get the app basically. Which, you know, is, is cool. It's just, like I said, it unfortunately hasn't had a ton of adoption.
3:16:04Vlad CosteaOh yeah, but people are lazy. They want apps. They just want to push one button and open that and be like, yeah.
3:16:12Karl KrederYeah. We always seem to be like fighting the lazy here, unfortunately.
3:16:18Vlad CosteaI've also noticed that for this wallet, there is no initial setup. Like it did not ask me to write my seed phrase or whatever.
3:16:27Karl KrederYeah. So, so that actually, if you go to settings, you can export your seed phrase. It has what we would call like incremental onboarding. So as you sort of get more Quai in that wallet, we teach you more about seed phrases and backups, but really Blip isn't meant to be, you know, cold storage. It's meant to be similar to the cash you keep in your wallet. So in the same way that for, you know, a couple hundred bucks, you drop your wallet, you lose a couple hundred bucks, it's like the same conceptual manifestation and Blip.
3:17:06So like, is the seed phrase important? Yes, but you know, it's only important when you have a lot of money on it. When it's a small amount of money and you're just using it transactionally or tipping or whatever it may be. You know, we deal with the concept of losing our wallets all the time, right? You lose your smartphone, you lost a thousand bucks, right? You lose your wallet, maybe you lose a couple hundred bucks. So basically, we have an incremental onboarding idea with Blip, which is we don't force you through that process until you need to go through it.
3:17:40Vlad CosteaWell, how does this work in the background? Like, is it SPV? Am I connected to someone's node? Or does this download information about the blockchain and display it in real time?
3:17:53Karl KrederThis is a great question. So presently in Blip, it is communicating to a set of RPCs. And those RPCs are trusted. So they're just like programmed in endpoints. But we have support in our nodes for allowing node operators to enable NiPoPoWs, so non-interactive proofs of proof of work. And any of the RPC requests, if they enable it, will go out with a NiPoPoW. Okay? Are you familiar with NiPoPoWs at all?
3:18:37Vlad CosteaNot really.
3:18:39Karl KrederOkay, so let's say I wanted to prove to you without you having a node that I sent you a Bitcoin. How could I do it?
3:18:57Vlad CosteaWell, it depends. Are we on different parts of the world or are we face to face?
3:19:05Karl KrederSo it doesn't matter in terms of me like showing you something. The question is, if I'm face to face with you, or if I'm in different parts of the world, is there information I could provide without you having to go to a third party trusted service or having to synchronize a node? So I guess this is kind of like an SPV concept, but it's a concise SPV concept. So maybe the answer to the question is, if I wanted to prove that I sent you a Bitcoin, what I would do is I would show you the transaction, I would show you a Merkle path, that Merkle path would then attach that transaction to a Merkle root of transactions that are in that block, right?
3:19:45I would then show you a certain number of blocks on top of that block, and if say that was six Bitcoin blocks of difficulty deep, you'd be like, cool, you sent me a transaction, right? So basically, we have that concept in our system, but because we have these structure of super blocks, and we actually have also a structure of interlinks, what we call them, they're like super, super blocks. They're blocks which aren't formed into a chain, but we create a hash-linked list that allows us to create a compressed representation of work that's been done.
3:20:25It's called an interlink. This is NiPoPoWs, like at its heart. But what you can do is, no matter how far back in time you go, and no matter what question you can ask, I can simply respond with three to 10 kilobytes of data. And you will know that to be true, if that makes sense. So the reason that we did that is one, we can because we're a proof of work network, all this nice hierarchical interlinking structure. But the other one is because once we upgrade our SDK, it means anybody that's using it, using the SDK will naturally get NiPoPoW verification.
3:21:04That means the app developer doesn't have to worry about this. They just import, you know, quais.js, and quais figures out how to route the request, as well as validate the proof that's returned with the response. But the implication is that every data consumer becomes agnostic of the data provider. And you don't have to do this by making everyone run an SPV, you can just do it using NiPoPoWs. So if you use like phones that try to do like clienting, they're actually quite cumbersome and take up a lot of energy.
3:21:35And then when you sort of turn it on again, it takes a long time to resync into the chain. But if you use the concept of NiPoPoWs, every one of the nodes in the network becomes a data provider. So then you don't have the same pinch points that you have with the currently existing blockchain systems. So if I use Bitcoin, and I'm not running my own node, I'm sending my transaction to like blockchain.com. Or I'm sending my transaction through Coinbase, I'm sending my transaction through an indexing provider.
3:22:08So we have this like big set of nodes. But if I'm not in that node set, there's like the set of maybe half a dozen dozen nodes where everything actually goes through both data going out in transactions going in. And it's because of the expense associated with doing SPV on device. So by using NiPoPoWs and compressed NiPoPoWs, we can do it much more efficiently by just returning the proof with the request rather than forcing the endpoint to run a node.
3:22:48Vlad CosteaThat's interesting. It's a good way to do SPV.
3:22:53Karl KrederYeah, it's an alternative to it's a much lighter alternative to the client running an SPV, but it gets you the same type or better security guarantees without like having to do that resource and syncing. And then the other thing that it eventually gets to is we believe that will actually create a market for requests, where people will pay either in hash or Qits for data from the nodes. So then the nodes can actually get directly paid, because all of the nodes are equally valuable to the end user.
3:23:31Right? The reason they don't go to a node is because they can't trust the node. So they go to blockchain.com, and they look at the certificate that says blockchain.com says this is true, and I trust blockchain.com. Right? So because of that, the app developer is then paying for the blockchain.com API, right, which is great, but that revenue can't be shared across the total node runner set.
3:23:55So the idea here being is that we want to create a competitive market for data providing, and NiPoPoWs creates a way to do that. And it also makes it so that you don't bottleneck the ingress of transactions in the network or the egress of data. Every node in the network becomes a data provider, they become an ingress point, and they can get paid to do it.
3:24:19Vlad CosteaYeah, that's a nice way of doing it. I just want to remark that since I added the QR code, people have been sending me small amounts, like five and ten Quai. Thank you very much for that. Someone in the chat said that this is how it used to be in Bitcoin many years ago.
3:24:37Karl KrederYeah, yeah, back in the day, and then Doge came in. Doge did this at one point, right?
3:24:43Vlad CosteaIt was popular with streamers.
3:24:46Karl KrederDoge's kick for a while was really the tipping crypto, right?
3:24:52Vlad CosteaBecause you could send such ridiculously high amounts, for example, 10,000 Doge or whatever, and it was less than a cent per Doge.
3:25:07Karl KrederUltimately, I guess, when we talk about the different use cases of the different chains, I believe that if you can function as money, that is technically the hardest economic use case. So I don't prescribe to the concept of running a game and putting every sort of in-game activity on chain. I don't think that that is economically useful nor sustainable, right? Like you jumping in a game or trading a sword, maybe trading a sword, but maybe that's an economically relevant use case. But the very, very high frequency thing is if you actually try to do an analysis of the economic demand for economic transactions worldwide, I haven't seen a good analysis of it, but I think it's around 100,000 TPS.
3:26:00If you're talking about like the internet connected world. So if you can function at that level, that's like the hardest thing to function as, which means that you can be used for anything that's above value, if that makes sense. So like if you can be money, then you're a better system for doing derivatives, you're a better system for doing stocks, you're a better system for doing anything that's built upon money. If you can be money is kind of how I see things. So the whole goal is if you can crack the decentralized proof-of-work system at scale that can function as money, you're actually the natural incumbent for everything else to be built on top of that.
3:26:48Vlad CosteaYeah, it's really interesting the way that you approach this. I have to say, I do like it, even though it seems like I wish this was released like five, six, seven years ago.
3:27:04Karl KrederI wish it was released five, six, seven years ago, man.
3:27:12Vlad CosteaNow it seems like even for payments, there are coins like Litecoin which have gathered a very large network effect. The price of Litecoin is not interesting, but the amount of merchant adoption, that's really something else. Like how do you imagine Quai getting integrated into this payment system?
3:27:35Karl KrederWell, I think there's some advantages to Quai in terms of being able to be integrated. So one of the unique things that I understand because of building the Lattice1 and developing that hardware wallet and specifically the smart cards on the hardware wallet is, I understand how the interchange networks operate, not just from a business standpoint, but also from a technical standpoint down to the packages that go between the cards and the terminals called APDUs.
3:28:20And my ultimate goal with crypto is to get it such that the interchange providers start supporting native crypto transactions. So not transactions that are going through a custodian holding Bitcoin and selling it every time you make a transaction for USD and then crediting the merchant, but one that actually just broadcasts the transaction to the blockchain. And so the interesting thing is we have a strategy which allows you to create an applet that would be compatible with currently existing interchange networks to directly do crypto transactions. And we have been shopping that around to a number of the interchange providers.
3:29:13And the goal being is that they initially let us do a pilot that would allow us to just set up service that replaces the issuer interface with just sort of a SaaS service that goes from Visa to blockchain. So it's an issuer interface that's just a proxy for the blockchain. So there's no custodian, there's no nothing, it's just this is the definition of how to interface with the Visa network, this is the definition of how to interface with the blockchain, and you're just putting a shim in between those two.
3:29:47And the thing that's interesting is they have a set of requirements to make that possible. And I think that only Quai can meet those requirements. Maybe Kaspa could too, but there's very few blockchains that would be able to meet the requirements that are required by the interchanges to use crypto directly through their rails, because they have stringent requirements on the type and composition of a message, as well as the time that they have to get a confirmation in. So we are hopeful, and that there's 18 different interchange providers and a number of issuing banks that will be able to find a partner and get the first demonstration of direct crypto from point of sale.
3:30:44Vlad CosteaBut I do have to ask you, you have so many transactions per second. What kind of internet do you need to be able to process all of this? And is there any, I guess, category of people around the world who are basically not able to validate fast enough? They're not able to stay in sync?
3:31:08Karl KrederSo the cool part about this hierarchy is that as the number of shards increases, the node requirement only increases very marginally. And the reason for this is if you're, say, a single shard participant, and here, let me let me share that screen again. That went on visualizer. Yeah. So if you're a single shard participant, see here, right, so you got prime region and zone. So there's four zone chains running. So what I can do as a participant is I can be on just this one zone chain.
3:32:06Obviously, I have to run region and prime because I'm sort of merge mining the headers, right? But those are just header chains. They're not state chains. So these chains up here are very inexpensive. But as long as I run header chains on the other networks, I don't have to run state chains, because what's happening is every block that's coming is a proof of work about that shards attestation to the advancement of that chain. So if I want to be a single shard participant, I can, which then means even as the number of shards increases, my requirements only very slightly increase with each shard because I'm just running an additional header chain as it happens.
3:32:56And header chains are very inexpensive to operate. So when we look at this setup right here, this is only a two by two setup. But when we went to 7000, or it was like 7400 TPS, and that was in a three by three, the requirements for the single slice nodes, or the single shard nodes was only still that Ethereum validator requirement. And the internet connection is still only around 10 megabits per second. So the cool part about sharding is that you can scale and you can do it in a way that doesn't require everyone to scale their participation requirements with the aggregate network requirements, if that makes sense.
3:33:52Vlad CosteaWhy do you think, so obviously you have more transactions per second and a design which seems more fleshed out than Kaspa, but why do you think that it was Kaspa that got more attention early on? And is the coin that's in top 100 on CoinMarketCap?
3:34:12Karl KrederI think it's the time that they launched, right? So I think Kaspa launched two years earlier than we did. So, you know, with proof of work systems, I think there's an adoption arc, there's a distribution mining awareness, like thing that happens. And I think that just kind of naturally takes place with time. So, you know, they certainly got to market earlier than us. But they also did it with a system that they were sort of upgrading. Maybe that was like a bad idea from our standpoint, but we wanted to get to market with a system we didn't really have to upgrade.
3:34:51I guess it was one of our design considerations. And we're kind of glad we did it, too, in some ways, because not only are you sort of violating the point of proof of work with continual, meaningful economic changes to the system that aren't meant to happen, right? I mean, the point of proof of work is to be static. It's to be a protocol. It's to know what the rules are and the rules aren't changing. So, you know, that was obviously one of the drivers. But the other thing that retrospectively I'm very happy about is it's much, much harder to work on a live system with an active community than it's to work on a pre-launch system because the iteration speed is so much higher.
3:35:35We could sort of, instead of having to think about migrations and backwards compatibility across a fork, we're just like, oh, we have a new idea. Delete the database and start it again. It's a much faster iteration cycle and test cycle that you could get doing this. And I don't think if we were live, we would have been able to get to this degree of complexity and endpoint in the same time period. If you look at Ethereum right now, on their roadmap, because of how developed they are and how many different clients they have and how many different tools they have, their upgrade cycle is one year.
3:36:17So they can't really go faster than a one-year upgrade cycle, but they have this extensive and aggressive roadmap of how to scale the L1. So I'm glad we didn't get ourselves into that situation.
3:36:32Vlad CosteaIt's definitely interesting because Kaspa and Alephium were the top, most profitable coins for, I guess, last year for the most part. I remember a news article that got published with someone who went to a hotel room and allegedly made like 5k, staying there and mining, with the electricity he got from the hotel.
3:36:59Karl KrederThat's hilarious. That's pretty funny.
3:36:59Vlad CosteaYou guys launched in the middle of a bull market. This should be bullish that you guys launched in 2025.
3:37:17Karl KrederYeah, I mean, like I say, I think with the arc of any proof-of-work coin, once you launch, because of mining emissions, there's an equilibrium price discovery that takes place. Where emissions reach equilibrium before you get real price discovery. I think we're in that phase right now. So, you know, it happened to Ethereum. It didn't happen to Kaspa because no one knew about Kaspa when it launched. They just started here and stayed here for a year before it started doing anything.
3:37:52Vlad CosteaDo you imagine that there's going to be a phase where someone makes ASIC miners? And what's going to happen next? Are you going to embrace it or are you going to change the algorithm?
3:38:07Karl KrederAgain, like our goal was to create a system and not really fool with it. Now we picked an algorithm that is GPU friendly as far as like we can see it. Unless there's sort of like a true existential threat to the system, I don't think we'd be changing the algo. So even if I guess ASICs did come online, that means they've kind of figured out how to deal with ProgPoW. ProgPoW is specifically designed to defend against them. So, you know, I don't know if it's good playing that cat and mouse game, right? So with ProgPoW, you have this concept of an epoch and you have this concept of, it's no overloaded term to call it a DAG, but there's basically like a manifest that you're searching through that's participating in a hash. And so it forces you to have that in high speed memory.
3:39:08So you see something that wants a lot of workers and you want a lot of access to high speed memory to be able to hash efficiently. So ProgPoW was kind of designed to make it such that if you were to build an ASIC to ProgPoW, it would look like a GPU and that, you know, manifest piece changes on epoch. So it changes. I don't want to say like what the parameters are because I don't remember them specifically on the top of the head, but it periodically changes and it also grows.
3:39:39And the point of that is to defend against ASICs and make it really just a GPU friendly algorithm. But if ASICs came in at that point, I think you kind of just let it go, because I don't know if it's like a good thing to get in a cat-and-mouse sort of fight over and manage what isn't supposed to be managed system.
3:40:02Vlad CosteaIn this industry, we have the so-called trilemma, which refers to basically an adaptation of Zooko's triangle, and in the iteration that we commonly use, it refers to decentralization, to security, and what's the other one?
3:40:27Karl KrederScalability.
3:40:27Vlad CosteaI guess if I ask you to which extent do you believe that Quai has them, you're going to say that you solved the trilemma. But I'm going to ask you to present this in percentages, like from 1% to 100% to which extent is Quai decentralized? To which extent is it secure? And to which extent is it scalable?
3:40:59Karl KrederI think it is as scalable as you can make a globally decentralized system. I don't think there's any way to make it more scalable. In terms of the trilemma, I'll say one thing. First off, I'll say the trilemma isn't sort of a real proof. I would say it's a relatively sophomoric attempt at trying to describe a very complex problem. But if you try to describe the complex problem, you're really talking more along the lines of CAP theorem, maybe with a fourth point that involves consensus. So I would say it's better described as a tetralemma, but your three points would be consistency, data availability, fault tolerance, and consensus is kind of like what you're operating amongst.
3:42:03And if we put it in the reduced dimensionality of that trilemma, effectively what we're saying is we're not breaking the trilemma, but each of the parts of the hierarchy have a different set of characteristics, if that makes sense. So if you were to draw it and put a point in the triangle, right, like maybe a zone is here, a region is here. And like prime is over here, right? So it's actually, as you're going up, you're moving in the triangle. And so what we're kind of saying is we think that we're getting the best of all worlds because if we think about a zone, a zone is slightly more centralized because the participant set is a subset of the total participant set, right?
3:42:59And it's only with the dominant blocks that we go from a subset to a bigger subset to the entire set, right? So there's, I'm not going to argue that a subsetted system is instantaneously as secure as a non-subsetted system. But I will argue that the amount of economic activity that I can put on that subsetted system far exceeds what I can put on the non-subsetted system. And with that statement, I can say, I ideally at scale then can get more revenue and more miner security than a non-subsetted system.
3:43:46So it's all about trade-offs. But I think that the trade-offs here are incredibly mild, practically speaking. Because if, for example, I want to buy a Spindrift, which are great if you're in America, by the way, but you're not, but beautiful if you ever are here. Sparkling water that's just sparkling water and juice, highly recommend.
3:44:16If I want to buy a Spindrift, my requirements are different than if I want to pay for an oil tanker. They're different requirements. But because in Quai, we have this tiered system, you can use it for both. So when we talk about finalization in a proof-of-work blockchain, you're really talking about two phenomena. You're talking about what we would call statistical finality and economic finality. Statistical finality occurs when we know the majority of hashrate is voting on our transaction. Economic finality occurs when a economically motivated attacker would no longer be profitable to reverse my transaction.
3:45:00So if I'm trying to buy a Spindrift and that Spindrift is $1, my most important factor here is that I just get it in a block. And the second that I have statistical finality, which we get very fast because we use the workshares, I also have economic finality and I can walk out of the store because I know that the store knows that I'm not going to try to do a private mining attack, which would be more costly to reverse that purchase of a Spindrift. But if I'm, say, trying to offload oil or put oil on a tanker and let it leave port, and you're going to pay me, that's a different requirement.
3:45:41I might need to send the transaction and kind of let that sit for 24 hours before I clear that ship off port until I know it's no longer economically viable to try to reverse that spend. So the idea with the hierarchy is that it allows you to get all of the properties you may want with the transaction, and it doesn't even change the requirement in terms of what economic finality means. So in a single chain system, economic finality is simply still when it becomes unprofitable to try to reverse the transaction through a private mining attack, and that has to do with the size of the transaction.
3:46:19In this situation, it also has to do with the size of the transaction, and eventually it's also a single chain that's attesting to that transaction. So we basically aren't sacrificing anything because we have the main chain, but we're gaining a ton of stuff because we have everything below it. But eventually, it's still that one main chain. So we don't lose that property. That property is still there. We just gain a bunch of properties. So as far as trade-offs go that I see, the only trade-off in this system that's bad is that settlement time, meaning in a high number of shards, you might have to wait up to 90 minutes to be able to re-spend money I send you, meaning you still know in five to eight seconds that you're going to get it.
3:47:06You can still get economic finality after, say, some number of blocks in terms of no longer being profitable to privately mine, but your ability to take that money that's coming to you and re-reference it to be 90 minutes. That's really the fundamental trade-off that exists here, and nothing else, I think, is of note, I would say. So the security is as good or better than any other proof-of-work system of similar hash. The scalability is, I think, as good as you can make the system, given the physical governing requirements of having a decentralized network, i.e. having many nodes distributed around the world, and that parameter is dictated by simply the hard drive times, the internet, wires, the speed of light, and computing power of the nodes. What was the other one?
3:48:09Vlad CosteaDecentralization.
3:48:11Karl KrederDecentralization, I think, is very, very high because we also have a ton of other mechanisms in the system that's trying to better optimize participation in an economically meaningful way. So I would say that in terms of how we've constructed markets, this is a construction which will incentivize the greatest amount of decentralization of any system. So I think we actually win that one in spades. There were two design criterias of this system. The one design criteria was Visa, MasterCard, and my mother. So what can she deal with?
3:48:56What's the transaction threshold? What's the throughput requirement? What's the transaction processing speed? That's like one. And then the fourth requirement was decentralization. So everything that we've been thinking about is how to create markets that incentivize decentralization. So when we think about the system, we're actually thinking about it as a series of resource markets. And we're always wanting to design efficient auction mechanisms for those resources. And if we design efficient auction mechanisms, we know that we always come to the point where supply meets demand, and you maximize aggregate economic surplus.
3:49:35If I can maximize aggregate economic surplus in every one of those points, we know that we're allowing the market to choose optimality of every characteristic of the system. So really what Quai is, is like we've tried to design a system that has as few presumptions as possible, and as many markets as possible, and both the outcome of the system and the outcome of the markets is determined by the participants. If that is, you know, how many shards there are, if that is, you know, what the emissions curve looks like, if that is the pricing of Qi, all of that is a market design mechanism. We're trying to be as non-prescriptive as possible in simply design markets. And we come back to this discussion about the nodes in Bitcoin and being able to run on Raspberry Pi.
3:50:35So, like, that discussion shouldn't be... So, you know, I mentioned that concept of an economically meaningful thing. What we're trying to do with NiPoPoWs and with workshares, which we haven't covered yet, is make as many participants as possible be able to be economically meaningful. And that is what you get to decentralization, is because if I have to beg people to run nodes and they're not economically meaningful nodes, I kind of get decentralization in name, but not in practice. But if I make everybody, if I make the largest number of people economically meaningful as possible, I get the highest degree of decentralization.
3:51:22And that's really what we've been designing to do.
3:51:25Vlad CosteaI was going to ask you because it seems like a trend within the industry, and I did have someone named Hunter Beast on the show like a month ago, and he spoke about his BIP-360 on Bitcoin. Basically, he said that even within the United States of America, in government agencies, they're pushing people to give up on cryptography that may get cracked by quantum computing. And the push is being made to make the replacement before 2030, which is in five years. So this leads to the question is Quai optimized to be quantum resistant or is this part of the roadmap?
3:52:15Karl KrederYeah, great question. So there's two different types of quantum resistance or quantum vulnerabilities, I should say. The hash functions, as we understand them, are not quantum vulnerable. The things that are quantum vulnerable in these systems are the signature mechanisms, so specifically ECC. And then there's two sort of sub buckets of that. There's what we would call short term vulnerable and there's long term vulnerable. Long term vulnerable is something like early Bitcoin transactions that were pay-to-public-key and not pay-to-public-key-hash. Meaning if I like I know what that key is, that key has money in it. And if at any point in time I can figure out what that key is, I can spend it. That's a long term vulnerability because maybe it takes me a year on a quantum computer to still crack the key, but I can do it.
3:53:15And then there's what we would call short term vulnerable, which would be I know that you've broadcasted a transaction, but before that transaction is finalized, I will try to attack you through mutation. And that is that's what you'd call like short term vulnerable. Even if we become long term vulnerable, which wouldn't apply to Quai. It would apply to Bitcoin, but it wouldn't apply to Quai. Which honestly, as somebody who has a tighter background in material science, I think all the quantum computer is like utter bullshit. But let's just go with the premise that it isn't. You're going to be long term vulnerable way before you're short term vulnerable. So we're going to see it and you're going to have ample amounts of time to figure out how to deal with the short term vulnerability, which would just be adopting a new signature.
3:54:19So if it were to ever happen, you would add signature support and people would just naturally start using it. So I don't think it's really an area that we should focus on. One, because we don't have the long term vulnerability and two, it's not necessarily a controversial change, because if it becomes obvious that it's happening because long term vulnerable systems are being exploited, everyone and their mother is going to move over to things that aren't short term vulnerable. And you're going to have a huge delta in time between going from long term to short term.
3:54:55Vlad CosteaOkay, that's fair. I'm also happy that you explain the two types of vulnerability to quantum computing, as I guess most people are not aware of this and they confuse what can actually happen and what cannot happen. What else should I ask about Quai? Is there anything that I missed and is super obvious?
3:55:20Karl KrederI mean, we can dig in more on workshares because workshares are really quite interesting.
3:55:27Vlad CosteaOkay, what's a workshare?
3:55:30Karl KrederSo a workshare, and if we go to back to the visualizer and I'll just do it on mainnet so it's easier to see. So a workshare here are these blue blocks. So they're basically equivalent to what you might think of as a pool share, but they're recorded and paid by the protocol. So they are basically attempts at finding a block, but they're sub block difficulty. So if you are familiar with how you get paid in a pool, the way that the pool ascertains how much hashrate you're contributing is your grinder, your miner, as it's mining, it won't find a block, but it will find what they call a pool share.
3:56:31So the pool will give you something to work on. You'll start adding nonces to that and doing hashing with the proposed block plus the nonce, the proposed header and the nonce. And some of the time, maybe you don't find a Bitcoin block because you're probably not going to find a Bitcoin block with 60 leading zeros, but maybe you find one with 20 leading zeros. That represents statistically to the pool a certain amount of hashrate going in. So as you're mining and every time you find one of these like 20 zero results, you send it to the pool.
3:57:06And then the pool records these and can statistically ascertain how much hashrate you're really putting into the pool. They add up all the hashrate of all the participants. And then when the pool finds a block, they divvy it up pro rata. Now, what we're doing in Quai is we're allowing those instead of having to go through a pool, they're going into the system itself. So one net effect of that is that you won't ever have to join a pool in Quai Network to get frequent payouts.
3:57:39So the payouts are always going to be frequent enough that even if you only have a GPU and even if we have like the Bitcoin hashrate, you're still going to get paid like once a week. So there's no like variability in payment requirement to join a pool in Quai. You basically have a pool in protocol, if that makes sense.
3:58:05Vlad CosteaThat's interesting. So how does this serve the purpose of decentralization?
3:58:14Karl KrederSo it makes it so that you you don't have that variability in payout and it makes it so that the incentive, at least that portion of it, doesn't exist in terms of having to join a pool. There's like five other things that we've added on top of that, but that's like statement one about workshares. So it's basically like we have a pool in the protocol.
3:58:45Vlad CosteaYeah, that's cool. I think Decred has something similar.
3:58:52Karl KrederRight. So that's one use of workshares. Another use of workshares is it is actually how we get fast finalization. So I mentioned briefly about sort of populations and samples. If we only take samples of the hashrate by the blocks and we have five second blocks, it will take us 30 seconds to get six samples, which only gives us an 87 percent confidence interval that we're representing the majority of hash. But what we do in our system is we allow for workshares and in expectation, you'll get eight workshares per block.
3:59:29So we actually go over a 95 percent confidence interval on the average block in a single block in terms of where the majority of hashrate is pointing, which is what allows for faster confirmations. So that's the other use of workshares. They're valuable to the network in terms of achieving statistical finality in proof of work consensus.
4:00:00Vlad CosteaThis graph is so mesmerizing, isn't it? I mean, it's a lot more complex because it has different levels or layers with the shards compared to Bitcoin.
4:00:17Karl KrederYeah. Yeah, and then the the prime kind of chain here you can kind of think of is like your Bitcoin. But then and this technically exists in Bitcoin too at some level. All of this stuff like underneath is is there. The problem is you can't implement it because to implement it, you have to change the protocol. That's like the fundamental problem. So for you to like do this to Bitcoin, you gotta pay shares. But to pay shares, you got to change the reward structure. And like nobody's gonna do that. Right like that.
4:00:51That's like why you can't build this on Bitcoin. Like like that's actually the fundamental reason you can't put this directly on Bitcoin is because you can't get them to pay shares. That's it. That's like the whole reason you can't do it there.
4:01:05Vlad CosteaI was gonna ask you this. If you had a magic wand and you could make one or two changes to Bitcoin, what would they be?
4:01:21Karl KrederPay shares would would be the first one, absolutely. Because with paying shares, like this that's like how this all happens. Like if you can pay shares, like Quai Network is just recursive. So so our like our protocol is like we figured out how to do consensus. But the the lines of code that generate this graph and I'm on the wrong window. I keep getting confused which window I'm in because I'm sharing a window. But how we generate this graph and how we generate this two by two graph. Like this is the same code, it's literally it's literally the same like 800 lines of code doing this. So so like Quai isn't complex, Quai is very simple. The process of proof-of-work mining, if you give it the simple manifestation, is emergently complex. Right.
4:02:27So like all of this is coming from an incredibly simple rule, which is PoEM, which is minimize entropy and then like this structure emerges. Right, but it's it's from this concept of sub shares and and super blocks and sub blocks. So yeah, if you could if you could do Bitcoin, being able to create sub blocks is in and pay them is like the way that you start to enable I think that's like level one. And then I would say level two, I would add a new hash commitment to Bitcoin of the UTXO set. Because then like SPV stuff would become like dirt simple. And then level three, I would add interlinks because then NiPoPoWs become inexpensive. And I would, yeah, I'd probably do in that order.
4:03:34Vlad CosteaThat's really interesting. I haven't heard anyone present these three specifically.
4:03:43Karl KrederYeah, yeah, but that's that's like why we can't like just look like, so this exists when Bitcoin models it's just not captured. But we can't capture it because we can't change the payout incentives. Like but this is happening. It's just like these zone blocks are pools. The shares are shares to a pool. Right. And then they just have a one layer roll up to Bitcoin. But that's kind of what's happening. So like all of this is there. It just isn't in the protocol.
4:04:29Vlad CosteaI hope all of this is visible and the people who are watching can see for those of you who are listening on Spotify. I apologize in advance that you're not not gonna be able to get the visual experience. Maybe check YouTube.
4:04:47Karl KrederThe visual experience is pretty sweet. There's it. I mean, this is this is us just like throwing this together in in web JS and like in a couple days. But you know, we've we've like seen this in our minds by reading you know log files and miner files for years. So it was actually kind of fun to see it for the first time in a 3d visualization for us because you know, we've been sort of operating in the matrix. Right? So I think you're like looking at Linux just outputs and we're constructing these graphs that are mine as we like built this system. So it's it's pretty cool to see in 3d and and we could actually do more like with these simulations with more shards. But it blows up the computer when you try it like it blows up your browser. Right because this is this is happening like in your browser and there's only like so many resources and so fast that it kind of works as it's visualizing it. But the chain can do it.
4:05:48The browser is kind of limited to four shards.
4:05:53Vlad CosteaWhen we started this interview, I mentioned that there is a similarity between Quai and the concept that's called tree chains, which is more than a decade old. I think what tree chains did not have is sharding, which is something that came along later. But the concept was very similar with having a blockchain for every different use case with different degrees of security. Ones that are benefiting from the activity of more miners and others that maybe don't have as much activity. And yeah, I think this idea was never really tried. It was just presented as a description on some IRC chat. And maybe I'm not sure if it's on Bitcointalk.
4:06:44I got to look it up. But I was gonna ask you if you're right about it.
4:06:49Karl KrederYeah. Yeah. It was it was just briefly described. I I didn't know about tree chains just anecdotally. I was thinking about the problem of how to scale proof-of-work blockchains and one of my friends was like, do you want to try mushrooms? I'm like, sweet. And we didn't like take a ton. But we were talking about consensus and blockchains and scaling while we had like a little bit of mushroom in our system. And that's the origin of like this concept. You are right to point out that is very similar to tree chains. There's there's some differences and there's a lot of additional advancements that he didn't get in the first pass. But of the things that have been proposed, it is most akin to tree chains.
4:07:49Vlad CosteaIt seems like I found a description of tree chains on Bitcointalk. And it says a tree chain is a structure of more tightly coupled side chains. This structure, in theory, allows miners to mine that arbitrary variance without pooling scaling of the system far beyond the seven transactions per second without asking permission and other innovation at the edges, all with the same protections of the main Bitcoin network. But the huge caveat that the idea is still half-baked has no known SPV client support and is much more complicated than a vanilla Bitcoin.
4:08:31Karl KrederYeah, so I'm the only one. Yeah, it's related. You know, I would say the differences are the concept of sharding, the concept of being able to do consensus without having to run full state like that is sharding in some level. The concept of PoEM, which is how you actually resolve the effective leader problem that takes place here. So with PoEM, because the weight is as unique as a hash, you're guaranteed to never have a conflict. And then obviously being able to then make that happen without getting withholding attacks and being able to have these, you know, super blocks without creating withholding attacks. That's all things like he didn't think through that we had to you know get through. And then I would say the other difference is our concept is homogeneous security and homogeneous execution and the system. We don't want heterogeneous execution or heterogeneous security. We think it's better to just make the system that can scale that everybody could use. And if it's all the same, then no one has to know about it. Right. It just it just melts into the background. So like it you use it, it works. It's like TCP/IP, but you don't know you know what your IP address is or what subnet you're on.
4:09:59Vlad CosteaI think I sort of ran out of questions, which is weird. You said you had a very interesting Vitalik story before we started this. Can you divulge it?
4:10:16Karl KrederYeah. So early Bitcoin guy on Bitcointalk, like bought first bitcoins in 2012. I heard about it very briefly on NPR. I think in 2011, taking a year to buy it. It was either the first or second most costly decision in my life. That only being matched with me meeting Vitalik. So, you know, I'm in graduate school and I like Bitcoin. I'm telling everybody about Bitcoin. CoinMarketCap exists. There's maybe like 40 coins on it. Right. Maybe we're doing a little bit of shitcoining, trying to get more Bitcoin. But it's fun. You're on the forums. You're hearing about things like X11, which turned into Darkcoin which turned into Dash eventually. Things like Scrypt and Peercoin and Litecoin and all these different things. So that's kind of like where I am.
4:11:33I'm doing my PhD. I'm reading all the forums. I'm reading all the posts. I'm reading all the innovations. I'm doing a little code contribution on the side. And so like I'm aware of these like ICO projects that have like started to happen. I think one of the early ones was, I want to say it was like NXT or something. Like somebody somebody did the very first kind of like ICO was like burn your Bitcoin and get this new token. And then after that, someone's like, yeah. And then and then I think it was NXT that was like, give me your Bitcoin and you'll get this token. And so like I saw NXT and I didn't end up doing it. And I was super disappointed after the fact because I would have made a lot of money. And so I was like on the lookout for ICOs that had some legitimacy. And this one Ethereum popped up in my radar.
4:12:32I'm like, awesome. I'm gonna research it. Like I see the team. Okay. Smart contracts. I get it. Composability, cool, let's go. Like I'm gonna invest in this Ethereum ICO. And I'm like, I'm gonna throw a hundred Bitcoin into it. But like what whatever at the time, it was like, I don't know, fucking maybe like 20 grand, 10 grand. It wasn't a ton of money. And my wife was like, let's go on spring break to California. I was like, okay. And you're wondering now how this has anything to do with the story. And so we decide that we're going to go to spring break to California. So we're gonna get on a flight from Austin to San Francisco. And then we're gonna kind of like take a car tour, you know, down through you know LA and then like back up through Monterey on highway one and whatever. Right?
4:13:32So I get on this plane and it's me, my wife and one of her friends and there's three seats. And one of the three seats, that's assigned, is next to this alien looking 19 year old. Who at this point in history, nobody would recognize except me. And it was Vitalik. And my assigned seat was directly adjacent to his. So the likelihood of this happening, I'm not going to a crypto conference. Crypto conferences like effectively don't exist. I think actually why Vitalik was in Austin was it was like one of the first Bitcoin conferences was like held in Austin and he spoke there and then was like flying back to like SF. So I end up on the same plane with him with the adjacent assigned seat. And I did the odds of calculating this. It's something like one in 80 billion based off some like assumptions about frequency of flying and whatnot.
4:14:30So I'm like, that's Vitalik. I guess I'll take that seat and I'll sit next to Vitalik. So I'm like, are you this Ethereum guy, Vitalik? And he's like, yes. So we basically sit next to each other for this flight and I start talking to him and I asked him, you know, I'm looking to put money into the Ethereum ICO. Sell me on it. Right. And he was like, programmable money. I'm like, okay, what are the use cases of programmable money? And he goes through like a couple use cases, like a smart contract.
4:15:03I'm like, okay. What's a smart contract good for? He's like an escrow. I'm like, well, how do you oracleize the escrow? And he's like, well, you have to have an Oracle. Right. Or like oracle wasn't a word at the time. Like I was like, how do you determine the outcome of the escrow? Like, you have to have an oracle. So anyway, we talked for like an hour and a half of this flight. I'm like, okay, that's interesting. When he pops open his laptop and he starts working on Geth to get it to market. And this was spring break of like, I think 14. So it was before the the ICO, which didn't happen like two months and like a year and a half before Ethereum actually like went mainnet. And so I get off the plane and my wife goes, how was your flight? How was the guy? I'm like, it was cool. But that guy's way too autistic to ever succeed. I'm not gonna invest in his shit.
4:15:50Oh yeah. I would have had 2,000 Bitcoin when it went to market.
4:16:11Vlad CosteaThat's pretty crazy. But when you look at it from this point of view, like. NXT had much better chances because the programmers seemed more experienced. They were building stuff. They had a deck that was working. Like if you put all of these elements together, you would say, well, this one looks more for professional and this Ethereum thing is just a bunch of kids who have no idea what they're doing.
4:16:40Karl KrederOh, it wasn't even like those factors. Like, honestly, it was I thought Vitalik was the only leadership. And when I talked to him, I'm like, there's no way this is gonna work out. What I failed to realize is that there were people like Joe Lubin and Gavin and Hoskinson. If I had like also sat next to those guys, they'd be like, okay this this could still work. But yeah, hindsight's 20/20, especially when you're looking at a pretty girl, right?
4:17:22Vlad CosteaI guess there's someone in the chat who wants me to ask you about why actually it's Qi and how this energy dollar can replace traditional USDT and can give the user the option to leave completely traditional. To completely leave, I guess he wanted to say traditional fiat system and ask how much it is worth right now and how to keep the price, etc. I don't know what he means.
4:17:52Karl KrederYeah, so Qi is the second token in the system, which is designed to basically just track the miners energy input costs. And it does that by making some assumptions about efficient markets and then setting markets up between Quai and Qi as well as an emissions policy of emitting Qi proportional to hash. So you know Qi we call an energy dollar. It's the one that has the cash like privacy properties. The way that it sort of tracks is twofold. The first mechanism in the system is miners get to choose the rewards between Quai and Qi. So they can say, I want to be rewarded in Quai or I want to be rewarded in Qi. And so the reason that they would choose Qi is because the market price of Qi is above the I'm sorry the yeah is above their production cost. And you know if that's the case, then they can sell it profitably at market, thus increasing the supply and the market price will tend to want to converge to their production. Now, the other mechanism in the system is the protocol offers an exchange rate between Quai and Qi. And with this exchange rate, you can burn Quai and mint Qi, or burn Qi and mint Quai. Now the question is, how do you ascertain what the offered exchange rate should be? And the way that you do that is by looking at the historic number of miner choices, as well as the speculator choices in terms of previous conversions to adjust the exchange rate until the market becomes neutral to the offered rate. When the market is neutral to the rate offered by the protocol, the protocol rate matches the external market rate of Quai, if that makes sense. So between those two mechanisms, you create a token Qi, which wants to always in equilibrium converge to the cost of electricity of the miners.
4:20:18Vlad CosteaOkay, is it shameful if I admit that I'm tired at this point?
4:20:25Karl KrederOh, no, that's that's fine.
4:20:27Vlad CosteaI know we were gonna go for seven hours, but I feel... I don't know. I don't know what to ask you next.
4:20:38Karl KrederLet's finish the Qi question. So the concept of Qi then is the price is supposed to represent the production cost, but it's based off the assumption of efficient markets. So as the markets are immature and they are small, the tracking or the stability of price is less good. But as the market matures, you would have the expectation that the movement in price of Qi, given those market mechanisms would become more and more stable over time. So I actually can kind of share that too, if you want to see, I had that pulled up somewhere. Yeah, so this is kind of like a dashboard that people look at, and this is the price kind of over the last 30 days. So you can see that it wants to trend to revert to a mean, but if there's a shock in either price or hashrate it takes a little bit of time to do that. But as the market becomes more mature, you would expect these oscillations to kind of get smaller and smaller. And really, the goal is that Qi converges to around $1.30, $1.40 per token.
4:22:01Vlad CosteaSo if it's like a stablecoin why would it be more than $1?
4:22:07Karl KrederBecause the market is not acting rationally. So the mechanism of how it prices is based off the behavior of miners and speculators. And so if there aren't an adequate number of sophisticated miners and speculators, you don't quickly get to equilibrium. So what you're seeing is sort of a very immature market. It's wanting to go back to an equilibrium price, but because the market is immature with not a high number of educated participants, you aren't seeing tight tracking. But as the market matures, you would expect to see tighter and tighter tracking and just kind of like a straight line across over time.
4:23:00Vlad CosteaAlso, I gotta give a shout out to the person who just sent me 400 Quai. Thank you very much for that.
4:23:09Karl KrederYeah, the Quai guys are generous. So that's that's Qi. I will say though the other interesting part sort of of that conversation that we didn't actually cover is workshares, because workshares have like six uses, but we've actually only covered two. And there's a couple other like pretty cool pieces in the system, if you're up to kind of covering those.
4:23:39Vlad CosteaSure, go ahead.
4:23:42Karl KrederSo one of the other things that workshares kind of does for us in this system is it allows us to not only make it so that miners can pay directly, even if they have very small hashrate but it also allows us to create something which we call a dpool. So one of the other reasons a miner might participate with a pool is because they don't want to run in the node. But the nodes, because a node has a kind of like an incremental cost to run, right? You have to have that you know Ethereum validator level hardware running.
4:24:24I don't really think it's the cost issue it's the complexity issue that causes people potentially not wanting to to run them, so it's much easier just to get a miner and point it at a pool. And those pools are centralized, which we don't like. So one of the things that people can now do, and this is pretty nascent, this is one of the things that this is kind of like X protocol that we have been developing after the launch, is something we call a dpool or a decentralized pool.
4:24:54So basically what you can do with shares is I can attach to a client in the network and they can provide me workshares or block proposed blocks. I can mine them. And then what they'll do is they will just inject their payout address every so often into the proposed block. And then based off of what is currently going to be a trusted system, but reporting of shares in this system, you'll be able to build reputation against the dpools to ascertain that they're only taking the fee that they're advertising.
4:25:39But what this means is that every node in the network not only can be this trustless RPC provider because of NiPoPoWs but they can actually be a trusted pool provider or a decentralized pool provider. They're not taking anybody's money, but you're trusting them to produce valid proposed blocks for you. And then based on those outcomes, you can rate each provider based off the data that goes on chain, if that makes sense. So we can actually make a system where we not only get rid of the pools from the requirement of getting payouts, but we can also get rid of the pools in terms of the requirements of the proposed blocks themselves.
4:26:19Okay, does that make sense?
4:26:24Vlad CosteaIt's pretty, I don't know, maybe you can compress that into a more elevator friendly explanation.
4:26:34Karl KrederRight now, if I want to point to a pool, I need to be like, I'm going to point to HeroMiners, and the reason I'm going to do that is because I trust HeroMiners. I trust F2Pool. I trust Foundry. What we're basically saying with workshares is they're giving us information that I can point at any node in the system and they can provide block proposals and I don't need to trust them. So I no longer need to point at a specific node. I can point at any node and there's a mechanism that they get paid for providing good block proposals to me and me to make money.
4:27:17So if I'm a miner and I don't want to run a node, I don't need to run a node and I don't need to join a centralized pool. I can just attach to any node in the network and I can mine against that node and I don't have to trust them to do it.
4:27:34Vlad CosteaThat sounds actually pretty good.
4:27:37Karl KrederSo it's another one of those design mechanisms of decentralization from an infrastructure standpoint, but also from an incentive standpoint, because now we've actually created two new revenue sources that nodes can potentially have that they weren't able to previously exploit, which is they can make money by providing proposed blocks to anybody who's willing wanting to mine, and they can make money by responding to RPC requests with NiPoPoWs. So you've created two new decentralized revenue mechanisms that any node can participate in and make money. Ideally, that's increasing the number of systemically important nodes in the system or economically important nodes in the system and preventing the creation of these pinch points where you take your decentralized system and you push them through a couple indexers.
4:28:35Vlad CosteaI just want to announce that we have basically reached 2000 viewers during this live stream which is really amazing. It's been four hours and a half almost and honestly it's not that bad given the limited time we had to promote this. I just want to ask you right now, Karl or Dr. K, let's find a special word for the people who have been listening up to this point so they can type that in the comments and we are going to know that they listened up to this point.
4:29:09What's the magic word?
4:29:11Karl KrederWorkshares.
4:29:13Vlad CosteaSo if you have been listening up to this point go to the comment section and type workshares and that's your proof of work basically. It's what demonstrates that you are a hero and you have been listening to this for four hours and a half or you just randomly skipped to this part and you want to flex. I don't know.
4:29:43Karl KrederWell if you do do it in the next five minutes or whatever it is I will make sure that the team goes through and tips you 20 Quai.
4:29:58Vlad CosteaOh, they would have to attach an address also, right?
4:30:02Karl KrederThey'll have to claim it. Yeah. Oh, no. No, they could they can just they can just do the the word will tip them on Kipper and then they go claim it on Kipper and there you go.
4:30:16Vlad CosteaWell, there are some comments on Twitter right now on the live stream, but I'm not sure if these will be visible later. There is Dr. West. There is Jonathan Downing. There is DoubleD.qi. DoubleD. There is Francisco Savia who says with the recent event of Qubic and XMR it makes sense to have workshares.
4:30:47Karl KrederWorkshares would actually fix XMR's problem by the way. So right now if you look at the on-chain data from Monero, they're getting periodically six block reorgs. So what Qubic is doing is they're basically running a selfish mining or a private mining attack and they're not putting any transactions in those blocks. So they don't have 51% of that hashrate currently of Monero. I think from what Shai produced it looked more like 37% to me. And obviously that's varying over time, but the max I saw them ever get to is about 37%.
4:31:31But even with 37%, they're having a massive impact on the liveness of Monero because they're effectively able to reorg six blocks. And with two minute block times, that means I could do a transaction on Monero. I could be 12 minutes past or 15 minutes past and my transaction vanishes off the chain after that fact. So in terms of functionally using it as payments and having a confirmation, it's really impinging on the usefulness of the liveness of transaction processing of the Monero network. But the reason for that is because they're only taking one sample of the hashrate every two blocks.
4:32:14So the variance of a minority miner that has this 35-37% are still going to win five blocks. If I look over a long time frame, I expect them to only win every third block on average. But in any given stint, they'll very often find five or six blocks in a row. And so they're using that to their advantage to compromise the liveness of the Monero network. But if Monero network had workshares, the blocks that are sort of lucky six in a row would not weigh the same as the blocks that aren't.
4:32:53So they wouldn't actually be able to execute this liveness attack unless they had truly 51% of the hashrate. So if any Monero listeners are there, implement workshares and this goes away.
4:33:08Vlad CosteaNot sure if they're gonna do it right away, but it does seem like a very good proposal.
4:33:16Karl KrederYeah, because if you did workshares, you could easily put, you know, 20 samples in a Monero block, you could put 50 samples in a Monero block, 50 shares. So there's no way that a sub-majority hashrate will get lucky over that sample size. They only can get lucky for like smaller sample sizes of like six or 10, which is what compromises liveness. But if you take a large enough set quickly, you know, Qubic could only compromise Monero if it indeed had a majority of hashrate.
4:33:51Vlad CosteaI was also thinking that another flaw in this design sometimes is that miners can censor, at least in the case of Bitcoin, right? Because they can see where the transactions come from and they can actively censor some transactions. And it would be nice if there was a way to encrypt the contents of the transaction so that there could be no discrimination. And the only criteria should be criterion because it's singular. The only one should be the transaction fee as opposed to the origin.
4:34:28Karl KrederInteresting. I haven't thought about how to encrypt the transaction in a way that the miner wouldn't be able to discriminate against it. I ultimately will say, though, as long as 100% of the miners don't discriminate against transactions, they still eventually get on chain. It just takes longer. Right. And an example of that would be Ethereum when they were doing OFAC compliant blocks. Yeah, blocks, right? Like 65 or 70% of the... I guess this happened during proof of work or proof of stake land on timeframes, but they got something like 60% of the network only proposing OFAC compliant blocks. But non-OFAC compliant transactions could still get through because, you know, 40% of the time, one of the other validators would still validate that block, which still processes the transaction.
4:35:42So, yeah, in terms of... There's one other thing that's actually interesting here with workshares, too. There's two other things that are interesting. One is the workshares are the mechanism which force the transactions to propagate. So when you're mining, you're creating these shares, but when you're creating the share, you're actually also including a set of transactions that you want to propagate in the mempool. And when you then share the share, you have to also include the transactions which you were trying to... That you were saying were valid in the mempool.
4:36:33And so this actually solves a problem that exists in basically all blockchain systems, which is known as the red balloon problem, which is you don't have an instantaneous incentive to share transactions. So if I'm a Bitcoin miner and I see a transaction come in, I would rather just take that fee myself and not share it with you. And that becomes especially pronounced if it's a very high transaction fee. If somebody put a transaction fee of a Bitcoin in as a miner, I would actually want to just hold on to that transaction until I win a block, not share it through the network.
4:37:09So the only reason they share it through the network is because if they didn't, there's kind of this like long tail of Bitcoin becomes less useful, but instantaneously, it's actually more profitable for everybody not to share transactions. So in Quai, because you're mining workshares, and those are blocks, if you're trying to include transaction in a block, you're also then including those transactions in the share, and for the share to be valid, you have to include the transactions with the share that you're trying to reference. So that encourages the transaction propagation around the network, which eliminates this non-incentive to propagate transactions that exist in every other network.
4:37:48That's another use of workshares in the system, because the workshares are also how you get paid. The majority of fee doesn't actually come from block, it comes from shares. And that's also how we get rid of selfish mining. So selfish mining effectively doesn't exist in Quai sub about 42% majority hash because the payments come through shares, and the mechanism of rewarding it that way is somewhat akin to FruitChains, which is a Nash equilibrium. So you can't make more money by withholding blocks or shares unless you have over about 42% of the hash.
4:38:30So yeah, also advantages of shares, it gets rid of selfish mining.
4:38:38Vlad CosteaI was going to ask you, since you mentioned that you're gonna basically improve the Quai Network for a couple more years before you think it will become final and, you know, just bug fixing and stuff. Where do you see Quai in the next couple of years, like in terms of both adoption and technical capability? Is there anything more that the laws of physics and proof of work can allow you to do? Or do you think that this is like the perfect balance, the sweet spot of what can be possible in terms of scaling?
4:39:15Karl KrederI think this is pretty much what you can do with scaling. So when we talk about like improvements to Quai, it's really like gonna be relatively minor stuff. I think there might be some tuning that happens with regards to that Qi control mechanism. But that would really just be kind of like refining how that's done. In terms of this system, I don't think you can design the system really much better than it's already designed.
4:39:52Yeah, or at least in protocol, right? So, for example... In Quai, we use something called access lists. And access lists basically allow parallelization of smart contract execution in an efficient way. So we don't have parallelized computation of the EVM done. But that parallelizing it would be X protocol, if that makes sense. So we wouldn't change anything about the protocol, we wouldn't change anything about the incentivizing. But we can change the client to make the client more efficient. We can increase the processing speed of a given client, a given resource.
4:40:38So there's some outstanding items like that, like we can make the data store in the client more efficient. We can make the EVM processing more efficient, but both of those efficiencies are X protocol. So you wouldn't see any changes, you wouldn't have to do the upgrades. But you could. So we, you know, that's some of the focus. Right now, we're really focused, though, on application development and making it so that more people can get Quai, more people can use Quai, and more people can spend Quai in different applications.
4:41:14So, you know, the first sort of arc of that is liquidity, right? So you're looking at exchange listings, you're looking at bridges, etc. And then I would say the second arc of that is really application use cases. And a lot of those will initially be digital and initially be crypto native. You know, Kipper is an example of that. There's another application, Trip Money, that's an example of that. The private messaging token is an example of that. And so there's sort of application development. And then after that, I think is really the arc of trying to get it to be adopted as money.
4:41:57So, you know, Blip, Pelagus, and trying to get merchant integrations, and ideally get one of the interchanges to allow us to create the first, you know, tap crypto payment that goes directly to a blockchain and doesn't get sort of custodian and sold by an intermediary. So that's kind of like how we think about the arc right now. So there's stages of it, and then each stage kind of gets you to the next one.
4:42:29Vlad CosteaThere's Dr. West at CryptoMaid1NUSA. I guess one is supposed to be I. And he says on Twitter, "workshares from Dr. K, our genius. He's a diplomat and a gentleman. Glad I sent you some Quai lol. Thanks for taking this time to interview. I've been in crypto for over seven years, and this has been my best day ever in blockchain. I learned more in five hours than I have overall. Thank you." This is why I'm doing this podcast.
4:43:05Karl KrederCheers, man. I appreciate the kind words. Yeah, I mean, I appreciate you taking the time to do the podcast too. You know, I think we've put a lot of thoughtful effort into really engineering of kind of how to do this. And I guess the hope is there's still enough people that care about peer-to-peer electronic cash to make it successful.
4:43:37Vlad CosteaI think that's sort of the biggest challenge. It's not the technicals, it's the culture. Because on one hand, you had to deal with 10 years of this. I'm not sure if it was 10 years, but anyway, almost 10 years of propaganda about blockchains not being able to scale. And then you had a bunch of experiments from a lot of people, including big blockers, Ethereum people, and those who prefer sharding, those who prefer pruning. We had MimbleWimble that launched and that one actually scales pretty well. But it's just that the overall sentiment was that we're going to need layer twos and then the layer twos are disappointing because they don't provide the same user experience and they have failures and they're just not exactly what we expected.
4:44:35And now we see a renaissance of layer ones, but it might be a bit too late and I hope it's not. I hope people still care about this. I hope they did not get brainwashed into believing that they should never spend their coins. And the whole purpose of this space is just to create speculative assets in which you invest. And you never use them for anything because really, that's not why many of us got into this space. And a lot of the guys I know, and I guess I'm also in that camp, we're hoping that we're going to get adoption and we're going to be able to buy our groceries with this and we're going to be able to do online commerce and bypass the fiat currencies and bypass Visa and PayPal and everything else that can censor us.
4:45:26Now, I guess it's only a matter of determining, first of all, which coins are going to be the successful ones because there is a steep competition there. And once we do discover which of the scalable ones are going to be successful, we've got to push them to get them to become adopted because it takes a lot of effort in this direction. And unfortunately, or maybe we can have more than five or five to 10 coins that are going to become widely accepted around the world. We can have regional trends where maybe there's going to be a lot of Kaspa in Argentina and there's going to be a lot of Quai in Brazil.
4:46:14I don't know. This might be the case. This can happen. It's just that I hope that it's not too late for us to get peer to peer electronic cash and be able to pay with it. And I also hope that you're going to be able to somehow tokenize Bitcoin for those of us who want to move BTC with higher speeds and maybe lower fees, too, even though I don't see how this benefits the Bitcoin network at large and how it pays the miners. But we're going to have to figure that out, too.
4:46:50Karl KrederRight. Yeah, no, I agree with all those things. I will just maybe kind of probe you on one thing. And I think you've alluded to this in two different contexts, one relative to Kaspa, but one relative to Grin. So I think one of your primary concepts when you say scaling is around state bloat. Right. So when you're referring to Grin, you're talking about cut through and the fact that cut through ideally minimizes state size for Grin-based transactions. And when you were talking about Kaspa, you said they know how to do a pruning of some sort. When I think about scaling and we're talking about state, state is the long end of the tail in terms of scale. The harder problem is not how to compress or trim.
4:47:51It's about the time it takes to get to consensus. And let me try to do sort of a rudimentary example. If I have proof of work Ethereum, or let's just go with the Quai block. I got a Quai block. And that block happens every five seconds. But it takes a half a second for it to propagate the network and it takes a half a second for it to process on average by the miners. I would intrinsically expect a 20 percent uncle rate in the system. Because when I find a block, it's going to take a second to propagate and get processed, which means during that second, everybody's still mining on the old block.
4:48:38And one out of five times, they'll find a block that competes with mine. So then every other time that happens, like I'll get uncle, but one out of five times an uncle will exist. And so we have this sort of 20 percent uncle rate. And the issue with proof of work systems is as uncle rates go up, your security and your finality goes down. So over sort of like a certain uncle rate, uncle rates become bad because you basically don't know that your blocks are final.
4:49:11Vlad CosteaWait, so what's the difference between uncle and orphaned?
4:49:18Karl KrederSo they're the same, except for an uncle gets referenced, an orphan does not. So in Bitcoin, there's nothing in the protocol that references orphaned blocks. When you reference an orphan block, we call it an uncle. So basically, the concept of uncling was introduced by a paper called GHOST, which actually Yonatan was one of the co-authors. But it's the greediest, heaviest observed subtree rule, which is basically saying in a system that has a higher uncle rate, when I'm trying to decide which sort of branch to pick, I should not only look at the single block, I should look at everything in that branch.
4:50:03So they add the weight of the orphans into the chain, and they're referred to as uncles. So when you reference those in, they're basically saying that this work was done, even though it might not have been done canonically on the tip. And they kind of showed that it is more secure than if you just have orphans. For Bitcoin, this isn't a problem because the orphan rate is so low because the block time is so long. But if you have very short block times, you can't have orphans, you have to have uncles.
4:50:36If you had orphans, your chain would become insecure very quickly. It would be exploitable by a sub 51% attacker, like almost instantaneously. So referencing orphans as uncles helps that problem for very tight block time games.
4:51:00Vlad CosteaOkay, that's interesting. So you started with this from the fact that I defined scaling in terms of storage. And for you, it's the time that's necessary to reach consensus.
4:51:16Karl KrederYes, so it's the time between when a block is produced, and when it is propagated and processed. That actually defines the bandwidth of scale. Like that's the number. So there's a couple physical constraints that go into that number. One is the propagation constraint. So if I look at the world, and I say I take the circumference of the world, it's like 250 milliseconds for light to circumnavigate. So I can't go around the world in less than 250 milliseconds. Now I could go halfway around the world in half that, so say 125.
4:51:59And that's the speed of light. Now, if we actually go across wires, and I have delays in switching, I have this delay of information propagating around the world. That's like this just intrinsic property of the network in light. So when I find a block, and I shout it to the world, let's say it perfectly propagates from my point out, it's still going to take 125 milliseconds to get around if it was a straight line, and it didn't have any hops. But in a peer-to-peer network, we're a mesh of nodes.
4:52:32So you have peers, I have peers, we have this networking layer. So for it to actually make it through a network that's not a straight line, it takes considerably longer than that. And that has to do with the total number of nodes in the network as well as their connectivity. So the larger number of nodes I have, the longer it takes for the message to get around the network. And that's one of the reasons it's like a very important discussion to talk about how many nodes should we have in the network.
4:53:02We want it to be enough to be decentralized both geographically, but if I have nodes in the network that aren't adding economic value, one of the things that they detract from is they make that propagation time much slower. I'll give you a material example. If we look at an idealized graph of nodes in Bitcoin, if I had say 10,000 nodes, I might expect a propagation time of two and a half seconds. If I reduce that to a thousand nodes, I might expect a propagation time of 400 milliseconds. Now for Bitcoin, it's not a big deal because you have these super long block times. But if I were to do that in a five second block, that becomes really important because I'm talking about the difference between a 50% orphan slash uncle rate or an 8% orphan or uncle rate.
4:53:55So obviously I want the lower of those two values, right?
4:54:03Vlad CosteaThat's actually a very valid point. And I think it's also going to be explored when I interview Yonatan next week.
4:54:11Karl KrederYeah. So with that, that's one constraint. Now the other constraint within that is the bandwidth. So right now we're just talking about ping times. If I just were to give you the shortest mesh as possible, how long would that take to propagate? There's also a bandwidth that underlies that too. So then if we talk about making super fat blocks, there's a time that they take to transmit, which then also adds into this delay time. So maybe I'm at like two and a half seconds, but you make a fatter and fatter block, the actual size of the block becomes more important because of the bandwidth limit.
4:54:45And suddenly maybe your two and a half seconds goes to 10 seconds, depending again on block size. So then the block size is a limiting factor in how much throughput you can have. Now then the next piece is the processing. So then if I have a very big block and I'm doing a bunch of transaction processing, there's going to be a time that it takes to process those transactions before I can start working on a new filled block. So that processing time has something to do with. So the real definition of scale in a blockchain, I think regardless of proof of work, proof of stake, it's from the moment in time that the block is created to when it is transmitted and processed.
4:55:38That time is directly related to how much TPS you can put through a single chain. And there's only so much that you can compress that without going outside of a globally distributed network. And I believe that number in the most optimized formats is about 400 milliseconds. If anyone ever tells me that something is happening faster than 400 milliseconds in a distributed system, you should just stop talking to them. Or at least I do, because they're full of shit. 400 milliseconds is like a physically bounded, like lower bound of if I have even a modicum of node participation, say about 100 nodes, like you just can't go below 400 milliseconds.
4:56:22And that's not giving like any time to processing or like any time to bandwidth transmission. So like anything under 400 milliseconds, if anybody talks about it, they're full of shit.
4:56:34Vlad CosteaI think you reminded me of a Romanian blockchain project that's called Elrond. I think they rebranded into MultiversX. Basically, they were saying that they can do interplanetary transactions and I was like, dude, if you send a radio signal with the speed of light between Earth and Mars, for example, it's going to take between 15 minutes to 30 or something.
4:57:04It doesn't work in the way that you described. There are the laws of physics in the way.
4:57:10Karl KrederSo if you slow down your consensus enough, you can do it, but your consensus has to be significantly longer than the magnitude of the time. So let's say we created another level in our blockchain and we called it the interplanetary, like right now we have prime region and zone. Let's say we created the interplanetary blockchain level and we said one of these blocks is only produced every six hours. Six hours relative to the 30 minutes it would take to go between like here and Mars is relatively small.
4:57:50So you could have like an interplanetary block time of six hours and still function under the propagation constraint of 30 minutes. But the time has to be significantly smaller than the block time, if that makes sense. Honestly, if you were saying it's a 30 minute constraint, you'd probably want to be on the order of like realistically as a baseline, like 3000 minutes in block expectation. So you could have an interplanetary block every 50 hours or something in expectation. So like you can do it, you just have to have inordinately slow block times for the interplanetary coordinating layer.
4:58:37Vlad CosteaThat's going to be a lot of fun.
4:58:40Karl KrederWell, so it's interesting when we think about this and we think about things like drivechains. The other the other property that happens when you talk about scaling chains back and forth is when one of the chains becomes certain about what's going on. So it's this interesting property of consensus that no form of consensus proof of work or proof of stake is deterministic. It's statistical. So like when I see a Bitcoin block, I don't know that we agree yet. Right. It takes some more time or some number of blocks before I know we're statistically or I think we're actually in the same chain on the same block.
4:59:28Right. So like I can't see one block. I got to see many. And that's why we say, OK, six blocks of depth, whatever it is. But I have to see more blocks before I'm like, OK, I think we're all in agreement. Right. And so when we talk about coordinating two different chains, like a drivechain, like we have to be damn sure that like we know the state of Bitcoin and we also have to be damn sure that we know the state of the drivechain. And I can't just say on average, I think we're good in six blocks.
4:59:57I need to say in the absolute worst case scenario, plus a factor of 10, I need to know that we're good. Because if I try to reference from Bitcoin to a drivechain or a drivechain to Bitcoin, and one of these things ends up not being true after that reference, you're fucked. And there's no way to resolve it, right? Because Bitcoin doesn't see the drivechain, the drivechain can see Bitcoin. But that means that if you have a reversion of state in the drivechain, Bitcoin doesn't care, right?
5:00:29Because like Bitcoin doesn't know about the drivechain. So one of the other reasons that like it makes it tricky to create sharded systems or side chains is because of this time property that it takes to get into consensus. One of the reasons that Quai can do this with so many chains, there's two reasons that Quai can do this with so many chains. One is that we reciprocally validate. So if Bitcoin validated the drivechain, then you could do this much faster. Because basically what you're saying is that if the drivechain rolls back, Bitcoin rolls back.
5:01:03And if Bitcoin rolls back, the drivechain rolls back. So if they ever have like an invalid reference, you get atomic resolution of the invalid reference. But that's a limitation to scale of Bitcoin is because Bitcoin doesn't see anything else, right? So in our system in prime, every time you find a prime block, it is only valid if it's also valid in region and zone. So prime is respecting region and zone, and zone and region are respecting prime. And because of that, we can create references that are almost instantaneous.
5:01:35And we know if, you know, something happens in terms of the fork, the system trims out exactly as much as it needs to trim out. And no invalid references remain, if that makes sense.
5:01:52Vlad CosteaYeah, it does. I was just looking. Go on. I'll let you know.
5:01:57Karl KrederYeah, no, I'm just I'm just trying to describe this. So I just, if you have a follow up, feel free to ask a question or anything.
5:02:09Vlad CosteaI just wanted to mention that we have been going for five hours and two minutes at this point.
5:02:16Karl KrederFair enough. You getting pretty tired? Are you hanging in there?
5:02:22Vlad CosteaI'm hanging. It's just that I'm starving. So I guess I miscalculated my resources for this one.
5:02:29Karl KrederYeah, fair enough.
5:02:31Vlad CosteaAnd I think it's rude to eat in front of the camera, but I was looking at the length of other episodes from the season and Zooko went for six hours. Peter Rizun, five hours, 24 minutes. What else? Vitalik went on for three hours and I sort of respect him for it because we agreed for two and then he went for another one seamlessly and he answered to all the questions.
5:03:07Daniel Krawisz, four hours. So you surpassed him. Ben Arc, seven hours, 23 minutes. My God, I had no idea. There are so many of them that are long. The record is eight hours, 21 minutes.
5:03:25Karl KrederWell, I don't think we're getting there, but I do want to prep you a little bit for Yonatan. So maybe I've described reasonably well this concept of what really defines the... Okay, so there's two pieces, right? So the first piece is just the data piece, which is like this concept of 400 milliseconds. Then you tack on processing and then you tack on consensus. And that time is the bandwidth. There's something about that time that intrinsically defines the bandwidth of the system. So what we're doing at Quai is with PoEM, we actually have a very interesting statement, which is in all other systems, consensus is only a statistical guarantee over time after this concept of consistency, meaning everybody's seen the same blocks.
5:04:22You still don't know that you all agree, even though you've seen the same data, right? With PoEM, what we can say is that we know once we've seen the same data that we've agreed. So we remove this tail of time to consensus completely from the system, and we're only bounded by that 400 millisecond propagation time.
5:04:49Vlad CosteaSorry, I was distracted looking at the stats. You said something about propagation time and then stopped.
5:04:58Karl KrederOkay, so when we look at the total bandwidth of the system, it's the total time really not to just propagate a block, process a block, it's also the time it takes us to get to consensus after we've seen the blocks, right? So what PoEM does that's very interesting is that instead of saying that only over time will we agree, it says if we've all seen the same data, we are guaranteed to agree. So in PoEM and in Quai, we don't have to wait these very long tails to do intershard referencing as long as we all see the same blocks, we'll always make the same choice, which means we're in consensus.
5:05:42So we don't have to add consensus time on top of propagation or processing time, we just have to say we're bounded by the processing and propagation times, and that's the bandwidth of the system. Consensus is not the bandwidth, like limiting factor, because consensus happens instantaneously with data propagation.
5:06:03Vlad CosteaThat makes sense.
5:06:07Karl KrederAnd so then the next step is the concept of statistical finality, and here's the issue that Kaspa has relative to Quai, which is this concept of sampling that we've talked about. When you create samples, the more independent samples you take, the better the sample is in terms of reducing uncertainty or variance. And every workshare in Quai is independent. The only dependent samples are the blocks. If we contrast that with Kaspa, every block that they put in their chain has many parent references. So every block they have is a codependent sample, which means their sample in terms of how valuable it is in terms of reducing the variance in the measured parameter of the hashrate is less.
5:07:02So they have to take more samples per unit time to get the same data. So the structure of their graph is demonstrably worse than the structure of Quai's graph. And I'll give you a very material outcome of that. If we were to take the same number of blocks produced, either as blocks or as workshares, I was doing an example and I think I was saying two samples per second, basically. And it was going to take us, I think, seven seconds to get to statistical finality, but it would take them something like 16. So because of the structure of their graph, it takes them twice as long to get to what would be considered finality.
5:07:58Vlad CosteaRight now, I was looking at the price of Quai to be sure about what happened. It seems like we pumped like 2% during this interview. I was curious about the effect.
5:08:12Karl KrederOh, yeah.
5:08:12Vlad CosteaThere's some volume going on. FYI, I'm not telling anyone to buy anything. This is purely a conversation about scaling proof of work. We're in it for the technology with gold chains, but technology.
5:08:35Karl KrederI'm not in it for the tech. I am in it for the ideology. I am just a technologist that's trying to build tools that accomplishes the societal outcomes that I'd like to see. To be perfectly frank, I am an ideologue.
5:08:57Vlad CosteaWould you still work on this if it went to zero? Would you try to fix it and bootstrap it and publish it once again?
5:09:07Karl KrederYes.
5:09:09Vlad CosteaThat's what it takes, right? That's also why I have faith in Bitcoin, because I know some of the guys who are still working to improve it have been around for a long time and they don't care about the money.
5:09:22Karl KrederYeah. It's about getting a system that accomplishes the societal outcomes. I'm not here because I like databases and distributed systems a bunch. I'm here because I like the societal outcomes of separating money from the state and operating on sound money. That's why I'm here. And I think to do that successfully, it's not just about scale. It's not just about the sort of soundness of the emissions. I think it's also about privacy. It's about creating a sustainable economic system in a proof of work setting, i.e. some mechanism to encourage velocity. I think all of those are required. And if I didn't think it was required, I wouldn't have spent the time working on it. But again, fundamentally, it's about creating a system that's so good that people will adopt it in their own self-interest, their economic interest.
5:10:32It's easier to use than other systems. It's faster. It's cheaper. It maintains or increases their purchasing power over time. It offers them privacy that other digital systems don't. It allows to participate in economies where they otherwise couldn't participate. It enfranchises the disenfranchised. That is what the system has to be able to do. And if you can't do all of those things, I think it's hard to see how we accomplish removing money from the state. That's why I'm here.
5:11:08Vlad CosteaWhy do you think it was not possible to do this with Bitcoin outside of the scaling issue?
5:11:15Karl KrederSo the scaling issue, I think, is the most pernicious. But second to that, I think there's just an issue with the velocity design. So if we look at history as examples, we were in a proof of work system when we were on the gold standard. That was a proof of work system. What is gold but a Neanderthal version of proof of work?
5:11:46The reason gold has value is because given how it's produced in supernovas and roughly evenly distributed throughout the Earth's crust, that I'm ensured no matter where I am in time, space or history, that it takes you roughly the same amount of energy to find, mine and refine gold. And that's why I ascribe a value to one ounce is because it's a manifest representation of the energy that went into producing.
5:12:16So we were on proof of work. And then the bankers creatively figured out how to leverage the fact that gold had issues with storeability, transportability and divisibility to move us into proof of stake.
5:12:31Vlad CosteaWell, yeah, but you can argue, for example, I've heard these stories about, for example, London banks in the 19th century under the gold standard. They knew exactly when there was going to be an audit. So they made sure that they have the gold in their coffers on the day of the audit. But otherwise, they were doing fractional reserve anyway.
5:12:55Karl KrederYeah, that's a great statement, which really means that the way that we get to where we want to go is self-custodianship.
5:13:04Vlad CosteaIt's very hard to prove anything. And I've even heard stories of banks basically taking the same gold bars from one to the other because they knew that on this day, this one is going to get audited. And then on the next day, it was going to be this one. They didn't keep the reserves that they were supposed to.
5:13:26Karl KrederWhy did the banks have the gold in the first place? It's because people don't want to store it at home, right? It's because the notes are easier to carry and circulate and divide than the bars.
5:13:38Vlad CosteaWell, that and the fact that some of them were earning interest on deposit.
5:13:46Karl KrederI would be very suspect of anybody that's offering the interest on a hard money asset, to be perfectly frank with you. But, okay, you know, loaning and credit is certainly a thing. But that's a different activity than the deposit activity. The activity deposits doesn't exist for loaning and credit. That's the activity of, you know, creating loans and credit.
5:14:15In the gold type setting, when you're talking about deposits, it's not because the depositors, there were many depositors that weren't getting interest. They were just there because the notes were more easily exchangeable and transferable than the gold. And so I guess the point is, if you can make it so that the transaction is decoupled from the investment aspect, you can still have fractional reserving and bank failures, but it only exists explicitly when people are taking risk.
5:14:54It isn't there when they're not. So if I have a $20 gold certificate, I don't have it because I'm trying to take risk against loans on the bank. I have it because it's more convenient than an ounce of gold.
5:15:12Vlad CosteaAnd I guess we also benefit from the advantage of being able to verify very cheaply that the transaction is valid, that the coins were received are part of the supply. I mean, how much does it cost you to verify the purity of gold, for example?
5:15:33Karl KrederYeah, no. So Bitcoin revolutionized that concept. It added a ton of properties that we didn't previously have in terms of being able to, you know, verify ownership and potentially reserves out of custodian. The problem, though, even with Bitcoin, if you're verifying reserves, you don't know liabilities. Right? So I can cryptographically verify what Bitcoin somebody holds, but I can't cryptographically verify how much they owe. Right. So no matter what, like you can't get yourself out of the situation with Bitcoin where you don't get some form of fractional reserving because you can cook the other side of the book.
5:16:13Vlad CosteaBut then again, nowadays we have all of these custodians and we have borrowing and lending and we have yield that you can earn by depositing your Bitcoin to some custodian. And we also have proposals for DeFi on Bitcoin.
5:16:30Karl KrederDo you do that, by the way, just out of curiosity?
5:16:30Vlad CosteaNo.
5:16:30Karl KrederDo you give anybody your Bitcoin to earn yield on?
5:16:37Vlad CosteaWell, hypothetically, if I had any Bitcoin, I would never give it to someone else, but I don't have any Bitcoin, fortunately. I don't have to stay up at night thinking about this.
5:16:51Karl KrederYeah. I personally, like, would never try to seek yield on it. There's other ways to do that with more transparent risk on assets.
5:17:01Vlad CosteaI mean, I don't have any Bitcoin. I have 633 Quai that I received tonight. Hope this is going to be worth something one day. I don't plan to spend it any time soon. But yeah, there is a market for this and there are people who want to do something like this. And a lot of it is happening on Ethereum. And I guess even more is happening with institutions that realize, OK, let's let's enable people to take loans against their Bitcoin and get leverage and whatever.
5:17:46Karl KrederYeah. Do you have any feelings on MSTR out of curiosity?
5:17:50Vlad CosteaI hate it. I feel like it's going to be the reason we end up in a bear market, which is going to be super brutal. And this time the stakes are higher. There are pension funds that invested in this kind of stuff. There are honest, innocent people who are going to be losing money because of this indirectly as they invested or they are subjected to some sort of insurance or pension that basically invested into this. Because they were maybe thinking it's safer than to buy Bitcoin or whatever, or institutionally it was easier for them to do it before the ETF.
5:18:37So they bought some MSTR. If MSTR makes it into the S&P 500, then millions more will be subjected to it, to the risk that it brings. I don't know. It's a giant scheme. I'm not going to say what kind of scheme because I'm not an economist. I'm not an expert. But basically, it's just a way to borrow more money against the hard asset that you have to purchase more of the hard asset. And when does this come to an end? When you run out of people to borrow your money.
5:19:15And what do you have to do at that point? Well, you have to sell what you have to pay back.
5:19:21Karl KrederYeah. People think it's going to work forever.
5:19:25Maybe we'll get lucky and the dollar will hyperinflate, so then the debts just go away.
5:19:33Vlad CosteaYeah. So I remember the pandemic. It was not so long ago and people were saying, oh my God, they're giving everyone $600. The money printer is going brrr. There's so much irrational spending. The inflation is going to be off the charts. Jack Dorsey got red pilled into thinking that there's going to be hyperinflation. And where is it, honestly? Like, I've been waiting for so long to see it. They were saying Weimar Republic level of inflation. Where is it? I gotta give credit to the central bankers. They knew exactly what they're doing.
5:20:37Karl KrederMaybe. I think so. So the thing that is under the surface that not everybody knows about was they eliminated the reserve requirement for banks in the US, and they allowed all of the banks to monetize their government debt at par. So when the Janet Yellen, or not, Elizabeth Warren constituency, like summarily executed Silvergate Bank, and then SVB, that caused a regional banking crisis. And the way they got the regional banking crisis to stop is they said, everybody who has government debt can deposit it in the Fed and we'll give you credit at face in the Fed. So, and then they eliminated the reserve requirement. So basically, there is no reserve requirement in US banks, and all debt they had, they could just put in the Fed and get full face value for even if it was like a 30 year dated note.
5:21:21So like that definitely happened. So I don't know if like we've escaped it. I just maybe the storm hasn't landed yet.
5:21:33Vlad CosteaYeah, I guess they're pretty good at doing a soft landing because we expected everything to collapse. But it hasn't. And we kept waiting and waiting saying, well, it's coming, we're gonna all get rich, there's gonna be a run to Bitcoin, people are going to be buying this. Never happened.
5:21:52Karl KrederWell, I don't know, man. Like I said, I, we might just not have waited long enough. We'll see. We'll see. We'll come back to this conversation in a year. See if things are still rosy with the dollar.
5:22:07Vlad CosteaI mean, if anything, if I had some disposable cash, which I don't have, I would buy more Bitcoin, which once again, I don't have as opposed to buying MSTR shares. I don't get it. But I do understand it's a thing in the US to have a 501k fund, and be able to allocate that into something. And yeah, some people were impatient, or some kind of ETF. So they put that into MicroStrategy. Some of them are happy. I don't blame them. It's just that I don't see this being sustainable.
5:22:45It's a huge risk. And I've seen Saylor present on stage two months ago. He was telling everyone to borrow a million dollars from people who may not ask you to pay back like I don't even know how that works. Like go to your parents and say, hey, mom, hey, dad, would you give me a million dollars so I can buy Bitcoin? Like, I don't care where it comes from. I don't care where middle class or whatever. And you've never seen that kind of money. Just give it to me.
5:23:17And then he was saying, well, you should leverage your future and basically take on debt against your Bitcoin and just keep going like this. And then you're gonna be one day you're gonna become a millionaire. For real.
5:23:32Karl KrederQuestion for you. It's kind of interesting, right within these ecosystems. When I started in Bitcoin, there was no like, entrenchment. Because basically, everybody that was there early on, was like fighting with it was countercultural. So like, it was about doing things differently and doing them with innovation. And there wasn't really like a sort of like, what's the right word for this?
5:24:17There wasn't heterodoxy. Right? Like everything was open to be discussed. Everyone was trying to sort of solve the problems together. Everyone was trying to figure out what you can do with it. And that was cool. And then at some point, like Bitcoin became, like, you know, I think it kind of peaked with like the scaling wars, but it became like more and more heterodoxy to it. And then there were like the upstart Ethereums and the Ethereums were like, we're gonna make smart contracts and fucking neon cats and rainbows.
5:24:50And, you know, we're gonna disrupt this and we're gonna, you know, make DAOs and this will solve governance and all these things. And there wasn't like a heterodoxy in Bitcoin or in Ethereum. But then, you know, probably after, you know, 2019, like a heterodoxy, like developed, and then if you were outside of that, you kind of got like, pushed out of the system. It's kind of interesting to me, I don't think you're a Bitcoin maximalist, but somehow you kind of like adjacently survive. And there's a couple people that kind of do this within the ecosystems, but I don't understand how it works.
5:25:29Vlad CosteaI am a Bitcoin maximalist. Like, I do want to maximize the ways in which I can use Bitcoin, right? I do want it to scale, I do want it to become fungible as a currency. To me, this is what maximalism is about, not telling people, well, you're a shitcoiner. Well, you should not waste your time doing this just by Bitcoin. Look at this chart. Your coin went to this point, it's 90% down against Bitcoin. I'm smart, you're stupid. I'm not really into that.
5:26:05Karl KrederYeah, well, yeah. So it's interesting, right? There's like a couple people that kind of will more openly discuss things that still haven't been like, kicked out of, you know, you might not be like center line, but you certainly have, you know, sway an audience where a lot of people just kind of get like thrown off the bus.
5:26:25Vlad CosteaBut you know, I'm like that guy who's telling jokes behind the church, you know, like telling jokes about the priest behind the church. And there are some guys who come along and they listen and they chuckle, but they can't go to the priest and tell him what just happened and they can't go home and repeat my behavior. So in some ways, I'm a jester in this space.
5:26:51Karl KrederYeah, it's interesting. But like, only a few people have kind of like gotten away with I didn't get away with doing that.
5:26:59Vlad CosteaNo, last year, it was horrible. I got more gray hairs because of that. People I thought were my friends turned on me. Why? Because they wanted to virtue signal on Twitter was really horrible. But I maybe the reason you think I got away with it is that I kept going and I didn't give up because it seemed like from the outside, it seemed like that was the end of it. It was like my career just came to an end. If I can call this a career, I don't know, it's just interviewing people on the Internet.
5:27:34But I was at a conference this year and I met someone else who is also a podcaster in Bitcoin and he's the virtue signaling type. He said, oh, I'm sorry that your show came to an end last year. And I was like, really? Is that really what you think? Like you stopped paying attention. I understand that. But you think I gave up because of that? And this was just like two weeks after I interviewed Vitalik and got a lot of views, probably more than he ever will on his shows that are so identical one to the other.
5:28:18Like, I don't even know why anyone would listen to something that tells you the same thing every week.
5:28:25Karl KrederYeah, no, that's that's kind of what I'm maybe getting to. I mean, it's just interesting if you try to have conversations that I believe are respectful of the purpose. And if I saw a way that you could directly couple this to Bitcoin, I would have done it. But like the problem is that without changing consensus, like you can't do it. So you at best, you can take an indirect coupling, not a direct one. But you know, you get kind of pushed out of these communities if you don't go along with the heterodoxy, which is both interesting and I think counterproductive for the communities themselves.
5:29:12Vlad CosteaI think I have some respect for the people who built the religion, because it does take effort and it does take memes to be able to do it. So they were able to build this community where everyone is thinking in the same way and reads the same books or pretends to read them at least. I think Chris DeRose, if you remember him, he made this very smart distinction between 50 page libertarians who maybe only read Bastiat's The Law and maybe On Liberty by John Stuart Mill. And that's it.
5:29:56And then you get people who read Mises and Rothbard and go through all the nuances of this school of thought. And not just Austrian economics, anything libertarian in general. Human Action or whatever. So basically you have to create a culture that converges those who only read 50 pages in their life and sort of organize their whole life around that. And those who spend most of their days discovering new ideas and trying to navigate this new information and make use of it. And creating a cult, even though it's sort of against the reason why anyone got here, because everyone who got into Bitcoin early on is a rebel in their own way. Being able to create a cult around this culture of being a rebel, that's interesting.
5:31:05The reason why I don't respect it is that it has these moving goalposts. Like they like to pretend something and then they change the rules. And they all agree with the new rules. And if you don't agree with the new rules, you get excommunicated like Andreas Antonopoulos. He was very instrumental, very useful for a while. And then the moment he did not want to change or update his views to the latest ones, he sort of got phased out, wasn't invited to conferences anymore. I guess also he gave up.
5:31:39Because he could fight back, but he didn't.
5:31:42Karl KrederYeah. I mean, he got kind of thrown off the bus. Roger got thrown off the bus.
5:31:50Vlad CosteaWith Roger, it's more obvious because he had this campaign of fighting against his first love, right? It was interesting that he was trying to call BCH the real Bitcoin. I understand why people still hate him for that. Maybe some people bought BCH and they're disappointed in the investment and the performance of it. I don't hate the existence of BCH. I think as an experiment, it's useful. As a social, how should I put it? So as a political movement, it was useful and maybe it made Bitcoin even more resilient because BCH was created at that point.
5:32:36And people saw that it's very difficult even with community members to have a successful fork that receives this aura of being the original network, the real Bitcoin. So I do understand why some people dislike him, but I think that they're also misguided or a bit brainwashed in regards to what the extent of his contributions to BCH. And they also ignore a lot of good stuff that happened there.
5:33:14Karl KrederI mean, I think he doesn't get enough credit for what he did for Bitcoin, regardless of how BCH came out.
5:33:20Vlad CosteaOh, no. And even those who were helped by him directly, like Mandrake, for example.
5:33:29Karl KrederYeah, it's pretty sad to see, man.
5:33:37Vlad CosteaBut as Paul Sztorc likes to say, it's a death cult. So at the end of the day, there is not going to be any true Scotsman left. So they either exterminate each other or they get to a point where they compromise on the principles, which I don't see happening, honestly. But it's an interesting monoculture that they were able to develop where they agree on stuff and at certain points it becomes very radical.
5:34:13But then you meet them in real life and you're like, how is this the same guy who looks so tough online?
5:34:18Karl KrederIt kind of sounds like the variation of the rules, right? So it kind of sounds like an actual cult, right? You end up going to the farm to find God and live a more connected life or something. And at some point it's decided that this guy's got it. He gets to fuck your wife, right?
5:34:50Vlad CosteaOh, yeah. I know that there was this joke circulating like four years ago that guys were saying, oh, please, Michael Saylor, impregnate my wife so I can raise a giga chad just like yourself.
5:35:06Karl KrederIt's literally the same thing, right? It is. Yeah. Yeah. Really insane. How did it come that Saylor, man? How did that happen? That he's like the guy, the Bitcoin guy now. He's like the most prominent face of it.
5:35:28Vlad CosteaIt's wild. It's very hard to convince me that he's not the CIA at this point, because he was always so good at infiltrating. He knew exactly like from the moment that he emerged. He was speaking in the memes that we sort of made up over the years, and then he was coming up with new philosophies and basically riffing on the same concepts. But by inserting different terms that he was making up by himself and kissing our asses and saying we're the cyber hornets and we're establishing a hive mind that's like the shield which protects Bitcoin and blah, blah, blah.
5:36:09And he went on every podcast out there, and what he accomplished by doing so was to establish himself as the leader, like the one who speaks on behalf of everyone else. Within one year of his involvement in Bitcoin, he was discussing with Elon Musk as the head of the North American Mining Council to find a compromise to get Bitcoin listed once again. At the Tesla shop. And that was the moment when I realized how much the community has fucked up, because this guy was able to just appoint himself as the leader of whatever and put himself at the head of the table.
5:36:57Why? Because he bought a bunch of Bitcoin. That's all he did. And interestingly, his move was to create quote unquote education, which means that he wants to tell people that Bitcoin is what he wants it to be.
5:37:16Karl KrederYeah, I always say you can't spell Michael Saylor without the letters CIA.
5:37:24Vlad CosteaYeah, you're right.
5:37:30Karl KrederPermissive license, use that however you want.
5:37:34Vlad CosteaIt's just that you have to go, it's M-I-C, so you have to go from C backwards to D-I, and then you have D-A. But yeah, also, if you are really into this rabbit hole, you can look at where the HQ of MicroStrategy is in Seattle. And it's just down the road from the CIA building.
5:38:02Karl KrederI mean, honestly, if you actually think about the rise of him within the system, now that you're talking about it more in terms of the time and everything, it's almost obvious what it is, isn't it?
5:38:16Vlad CosteaOh, if you look at where he went to high school, it was West Point. He trained to be a military pilot or something.
5:38:27Karl KrederOkay. He makes more sense, doesn't he?
5:38:31Vlad CosteaWhen you look at it like he went to this military high school and then he started his own company. What does MicroStrategy produce anyway? No one knows. You can find on their website, they're doing software for all of these government institutions, and they do have some three-letter agencies in there for which they made software.
5:39:00Karl KrederMakes even more sense, doesn't it?
5:39:03Vlad CosteaIt does. I mean, it's... I don't know. And there are people out there who say he's like the savior, and there are people who believe he's Satoshi Nakamoto. Came back and said, okay, let's buy more of this because I cannot access my wallet from the early days. But you know what's funny? There are a lot of communities in this space that believe that their leader is Satoshi Nakamoto. Kaspa people? Yes, Yonatan is Satoshi. Litecoin people? Charlie Lee is Satoshi. Zcash people? Zooko is Satoshi. And I'm pretty sure that there are more of them.
5:39:50Karl KrederI think it's Hal Finney, to be honest with you.
5:39:54Vlad CosteaNo, I'm not sold on Hal Finney.
5:39:56Karl KrederReally? Who do you think it is? And honestly, I actually think it was probably three people, not one, but...
5:40:05Vlad CosteaI would think it's Nick Szabo. That's my favorite. It's just that Hal Finney was so versed in writing C++ code and everything. And he also contributed to fixing some bugs. And the way that they were conversing, right? If you look at the cryptography mailing list and see how Hal Finney was replying to Satoshi, it was almost like a professor to student kind of relationship. Maybe it was on purpose, but I don't know. It's just that I'm not sold on Hal Finney having been Satoshi.
5:40:51Karl KrederYou think he would have written better code in the initial implementation?
5:40:57Vlad CosteaYeah. I mean, he's the guy who wrote the second version of PGP. And he was a very respected cypherpunk.
5:41:13Karl KrederYeah, but whoever it is, I don't think they have access to keys or they're dead.
5:41:21Vlad CosteaI guess we'll never know, unless there's gonna be some sort of move of funds. But I think if there's gonna be one, it will probably be caused by quantum computing, brute forcing.
5:41:38Karl KrederYeah, I mean, I would just put it this way. If I were Satoshi, even if I had access to those keys, I would have burned that a long, long time ago. Because I think he wanted to achieve something. And I think he would have recognized that those keys being in existence anywhere was a huge threat to Bitcoin. So I think he would have personally disposed of them. One, because the only reason to keep them around would be notoriety or to move the tokens, which would not be a good idea.
5:42:21So I think that he would have banished the keys is my read on it.
5:42:29Vlad CosteaSomething interesting that I was able to get on this podcast was during the interview with Martti Malmi, who was basically the first assistant to Satoshi. He discovered Bitcoin in April of 2009 and spent the next couple of years working on Bitcoin. He worked on the web page. He worked on writing tutorials and guides on how to use it that were published on Bitcoin.org. And then he wrote on the text that was published on the P2P Foundation when Bitcoin was announced there. And what Martti told me is that he does not believe that Satoshi was some sort of NSA cryptographer or CIA agent because he saw him struggling to configure WordPress.
5:43:23And he said that this small detail about the way in which Satoshi was working is enough of a reason for him to believe it was just a guy who was not very or extremely competent in what he was doing. He was not like a master coder or whatever.
5:43:45Karl KrederYeah, he was clearly like a little bit of a polymath that maybe had more... I personally think, looking at his code, he looks like more of an embedded controls person or an electrical engineer. And having formal background more closely related, it seems like he was using the tools of cryptography and using the tools of hash functions that those weren't where he existed. This is my take on it.
5:44:19Vlad CosteaBut once again, within the cult, there are people who believe that Satoshi was some sort of alien entity that came to our planet and decided to give us the perfect form of money, or that Satoshi was this time traveler who came and gave us, once again, the key to perfect money. And they also search for secret meanings in some parameters, like the 10 minute block time, the 210,000 block. Having the 21 million coin cap, like, sure. But there is this demo that has been circulating before Bitcoin launched.
5:45:00Satoshi reached out to the cypherpunks. And he had an early version, like the alpha version of Bitcoin from November of 2008, which he sent to Ray Dillinger and Hal Finney. And actually wrote about that. And I did a presentation last month at a conference. Basically, the block time in the original design was, if I'm not mistaken, 15 minutes. And the halvings were happening every 100,000 blocks or something. I got to look up my notes, but the parameters were very different. I'm pretty sure that they were arbitrary, because in that design, if you look at all of this input data, you're going to discover that the maximum supply of Bitcoin under those rules would have been 20 million.
5:46:00So I don't think there's anything about the number 21 that's special. It's just that Satoshi decided that the reward should be 50 Bitcoins per block. And there's going to be a halving every four years. And then the original design, he had 100 Bitcoins per block. And the halvings were going to happen more frequently.
5:46:25Karl KrederYeah, I think he was a man. I don't think he was an alien or anything. And there's a couple pieces where it's pretty obvious. Early on, he didn't even have heaviest chain rule. He only had longest chain rule, which in a variable difficulty setting, creates an insecure chain. So in the first year, they had to go from longest to heaviest. And there's other pieces, like the whole conversation with the block size. The guy wasn't perfect, and the one megabyte was never supposed to be a permanent fix.
5:47:01It was just to keep DDoSing from happening initially. And you look at that conversation, it was like, yeah, we'll put it in there for now and we'll figure it out later. Seems like an engineer, man.
5:47:13Vlad CosteaThere's also the first proposal to increase the block size, which was also witnessed by Satoshi. It was by Jeff Garzik. And he basically published a modified config file and said, you can just put this and you increase the block size to seven megabytes, if I'm not mistaken. Satoshi commented and said, don't run this because you're not going to be in consensus with the rest of the network. It's a bit too early for this. This was also an argument that was used by big blockers during the block size wars.
5:47:46Karl KrederYeah, no, I mean, even the economics and the discussion of economics, when he talked about the sort of inflationary subsidy going away, it was going to be made up by supplanting it with transactions. And I think looking at it and saying that the expectation that it was still going to be five TPS that was going to be able to offset the block subsidies was probably not what he was considering when he was looking at that. And he's like, yeah, we'll have a ton of transactions. So they'll pay enough money to keep the security up, not five.
5:48:25So that's my read on that as well in terms of I think he expected the throughput to go up to create that sustainable mechanism over time.
5:48:39Vlad CosteaI mean, when he was dismissed and this was in August, no, it was October, I think, of 2008 on the cryptography mailing list. It was James A. Donald, the first person who replied to him and said, well, this doesn't scale. And Satoshi did some very quick napkin math and said, well, if you can download the DVD today in this amount of time, assuming that Moore's law is going to continue and we're going to get faster computation, more RAM and also faster Internet, then we can easily scale this to Visa levels.
5:49:19And once again, this was an argument that was used by the big blockers during the scaling debate that Satoshi himself said that you can scale a Bitcoin with the current design, without layer twos, without anything else.
5:49:38Karl KrederYeah. Well, again, everything that I've seen about it, looking at the early code base and whatnot, I think the guy was an electrical engineer, but the point is he took an engineering mentality to it as well, which is, it's not perfect, but... Like, this is kind of good enough for now. We can't make things perfect. Like, let's go with it and we'll fix it in post, so to speak. We'll learn something and adjust, right? Whoever he is, he did a great thing. Hopefully, we end up on the right side of the outcome of digital money, is all I got to say about it, because...
5:50:32We end up in a custodian world. We end up in a world without privacy. We end up in a world of CBDCs. Like, this world's going to get pretty ugly, especially considering what we saw happen during the pandemic in places like Canada and places like the UK and places like China, too. It's horrific, both in terms of speech, coercing people to take, you know, shots that weren't tested, imprisoning them effectively through quarantine or hospitalization, not allowing doctors to practice medicine, right? So, you know, with our authority is completely untrustable.
5:51:12They're pernicious. They abuse it. And the ultimate sort of power in this world still stems from the control of the money system. So if we get the digital money and they can track, trace, confiscate it, they're going to have ultimate power. And I think at that point we're all kind of hooked.
5:51:34Vlad CosteaAnd the way that you presented it, it became quite depressing very fast. But yeah, we should be optimistic because we have this technology. It's just a matter of putting it into the hands of the masses, right? Proving its utility. Because, of course, there's a lot of speculation, which isn't always bad, right? It's just the process of price discovery, where you're trying to understand how much this is worth. And people decide based on the events or whatever their beliefs are, whether or not something is worth their money and just keeps going. What's really important in this space is for projects to have consistency, which is not something that you see because a lot of guys are here just to make some money.
5:52:31They launch something during a bull market. They make money off of it. They see the price collapsing afterwards. And instead of trying to rebuild it or trying to get users to get the value that they expected, they just move on to another project and so on and they repeat the cycle.
5:52:55Karl KrederYeah, I personally think that the industry, so to speak, I would say 99 plus percent of the stuff is just absolutely irredeemable. And probably 90% of participants aren't honest actors, to be perfectly frank. At one point in 2017, 18, 19, it was like, I was cool telling people I worked in crypto, but after 2021, I was like, fuck.
5:53:37I do software development.
5:53:42Vlad CosteaOh yeah, that's what I tell people when they ask me, well, what do you work? What's your job? I say I work in IT. IT, yeah. And they're like, okay. They don't ask any more questions. But if you tell them, oh, I work in Bitcoin, they're kind of like, oh, is it a good time to buy? Or they express their opinions based on what they heard on the news that happened. They're like, oh, you must be doing very good right now. Or, oh no, I heard the price went down.
5:54:09Karl KrederI heard Bitcoin got hacked.
5:54:11Vlad CosteaYeah, exactly. Like someone told me that it's going to zero.
5:54:20Karl KrederWell, cool, man. I appreciate you taking the time to dig into Quai. I mean, I think we're at six hours here.
5:54:32Vlad CosteaAlmost. We need five more minutes to get to six hours.
5:54:38Karl KrederWho's the next guy to beat? Was it 6:06 even? 6:02? How are we doing this?
5:54:46Vlad CosteaI think it's top 10 in terms of length.
5:54:50Karl KrederYeah. What do you think it is in terms of novelty?
5:54:57Vlad CosteaInteresting question. I don't think it's a fair ranking when I think about it. Because most people were more related to Bitcoin and trying to build something on Bitcoin. In terms of novelty, it is top five, at least this season. Because there are so many concepts that I was not familiar with. And I still don't think I understand them very well. Maybe I should try to look at a block explorer more to understand what's happening. And maybe run the software like a proper node. Or try to mine it to fully grasp what's really happening.
5:55:43Sharding is complicated. It's not easy to implement. I know that it was a big deal. It was sort of like the holy grail in Ethereum back in 2018. And then I think they gave up on it for a while. And that's why a bunch of quote unquote Ethereum killers that do sharding were launched. I'm not aware of any one of them that actually succeeded. But proof of work sharding, good luck with that. I think only Alephium, which is trying to do something Ethereum-like with smart contracts, is proof of work and does sharding. And then there is Quai, which launched this year. Alephium has been around for like three, four years. I'm not sure. I've interviewed them last month.
5:56:40But it is very interesting. And there's a lot of novelty. I'm happy that research gets pushed beyond what we think is conventional. And also some people, I remember very clearly that one of the arguments for Ethereum going proof of stake was that they can get transaction finality much faster.
5:57:07Karl KrederI disagree with that, by the way. I think proof of work is fundamentally more scalable than proof of stake.
5:57:13Vlad CosteaAnd your work proves it, right? Yeah. It's easier to just go proof of stake, I guess. There are far less factors to take into consideration, especially if you know that a majority of coins is still in the hands of the founders or whatever, and they're not going to become evil.
5:57:43Karl KrederYeah, it's actually like a harder consensus problem, though. So when we talk about that time we were talking about that dictates bandwidth, the interesting property about proof of work is that when I broadcast a block, I don't have to talk to you back and forth about that block if you use a deterministic mechanism for choosing a chain like PoEM. So I just broadcast it, and I know by the time you see the same blocks that I've seen that we will choose the same block. But we do that without communicating.
5:58:21So I just broadcast it, but I don't have to be like, hey, Vlad, did you pick A or B? I know that you picked A if A is heavier, and I also picked A if A is heavier. So we're good. So consensus actually happens when we start the blockchain at Genesis and we go around the rule. And we know that we all agree because we know the rule is deterministic, right? The problem with proof of stake is it makes that time to consensus worse because we have to talk about it.
5:58:47So I'm like, I have a proposal for what this block should be, and I'm the elected validator. Would you like to sign it? And then you have to sign it, and then you have to pass it on. Somebody else has to sign it, and we have to keep passing it around until we get a quorum. And then we got to broadcast that quorum of the proposed block and signatures. And then we're like, OK, we think we all agree, which is a lot longer process than me just throwing blocks into the universe, right?
5:59:12We don't have back and forth, and we don't have rounds of communication. So fundamentally, that characteristic makes proof of work more scalable than proof of stake. And it makes it more compatible with sharding, much more compatible with sharding than you could possibly do in a proof of stake setting.
5:59:33Vlad CosteaYeah, but I don't know. Maybe there are some memes, right? Because a lot of the narratives in this space are dictated by memes or rhetoric. It's never conversations like this for better and for worse, I guess, because I'm not sure who has six hours to listen to a podcast. That's sort of a luxury for most people. And if they do listen, I don't think that they use the knowledge that they acquired to try to influence the opinion of the masses. I think this podcast is very nerdy and niche, but in the best way possible.
6:00:15The people who listen actually build stuff or they do something meaningful. I wish I had more influencers listening to this and trying to take ideas and propagate them. It's just that it doesn't really happen based on my experience. Over the last six years since I started, I've started to understand who is really listening and who actually cares about this.
6:00:48I don't know exactly who, but I do meet people at conferences who tell me, oh, I've listened to this interview of yours and that's how I survey, sort of based on this assumption that it must be this bunch of nerds. And once again, people who go to conferences, they have a higher level of involvement in the space as compared to just the casuals who are only throwing money at projects and expect this person on the screen to make them rich.
6:01:23Karl KrederSo what you're saying is that I should become a cult leader?
6:01:27Vlad CosteaIf you really want your coin to pump, look at Charles Hoskinson. Seriously. Like, what does the guy do? He just takes funny pictures and gets on Twitter spaces, talks a lot, and then gets on YouTube. And he has like a two-sided personality. On one side, he talks about how wealthy and successful he is. On the other side, he talks about some futuristic mumbo-jumbo about his Cardano. And it works for some reason. I don't understand why. I mean, I'm not part of his target audience, I guess, but I've never looked at it and said, this guy is so smart, I really want to give money to whatever it is that he's doing to invest.
6:02:18That never happened to me, but it seems like it's super popular.
6:02:24Karl KrederYeah, it's interesting. Maybe I should just let my hair down and get a white robe and just be like, we'll save the world. Yeah, preach.
6:02:33Vlad CosteaThat's how it works, unfortunately. That's how it works. There's nothing you and I can do to change it. It's not like you can take everyone and keep them in school for a few years and then get them back and explain to them that they should also think for themselves and stuff like this.
6:02:55Karl KrederYeah. Or you get PA and people come in because they're like, ooh, look. And then you use that to build user bases and applications.
6:03:09Vlad CosteaIt's also painful to see that most of this space is not real. Like you deal with a lot of bots and a lot of marketing agencies that mobilize their armies. And you see some of the same kind of messaging across multiple projects, which makes you realize, OK, there is some common ground here, which is suspiciously similar. And then you see these people disappearing completely during the bear markets. And it makes you wonder.
6:03:41Yeah. And then the projects will say, well, we just reported we have a million transactions per day or something. And then you're like, so who are the users? Where are they? I never see them anywhere. And if there are a million transactions per day from real users, maybe that your coins price would not be correlated with the price of Bitcoin. Maybe you'd have real demand for it.
6:04:09Karl KrederRight. Yeah. I mean, it's interesting. Steam adoption of Bitcoin was like really cool. Even Silk Road stuff was kind of cool initially because it was basically serving a disenfranchised set of people or just something that didn't have rails. Right. And the reason that Microsoft was interested in using Bitcoin as payment on Steam back in the day was because they didn't have good credit card access in Southeast Asia. So they didn't have any way to get people to pay for games. Bitcoin was a better payment solution. And it was like a sad day when Steam was like, no one's using this anymore and we're no longer accepting it.
6:05:02Vlad CosteaI understand why, but it's so disappointing because there is a huge network effect and it's a very nice community on Steam. I think video gamers are some of the people who are most likely to understand Bitcoin because they deal with virtual currencies. And a lot of these virtual currencies right now aren't just gold in World of Warcraft or whatever. They are skins. They are all sorts of options within games that you can trade.
6:05:39And they understand both the concept of a virtual currency and of a virtual marketplace. So I wish more video gamers cared about Bitcoin. But coincidentally, I also edit and run the only video game magazine here in Romania. And we publish a new issue every year, sort of, sometimes two per year. But I do realize that in spite of my efforts to get them to understand Bitcoin, they're like, no, this is not for me. I just want something that I can spend on games or whatever.
6:06:23Karl KrederAll right. So, if Steam took it, they'd be like, I love Bitcoin. I can spend it on games.
6:06:27Vlad CosteaExactly. Yeah. Pretty sure. They would be like, oh, I just bought $10 of Bitcoin to buy this game and I forgot to buy the game. And now two years later, it's like $100 or something. This could happen. And they would be like, oh, I can buy six games with $100. This could be so cool. It's just that, I don't know, there was Steam, there was Expedia, there was Dell, there was Microsoft.
6:07:00A lot of these places were accepting Bitcoin. Also Tesla. Overstock.
6:07:04Karl KrederOverstock was one of the early ones.
6:07:06Vlad CosteaWhat's Overstock?
6:07:08Karl KrederOh, Overstock? That was Preston Byrne, I believe was his name. I think he might've been like the first, like the very first to accept Bitcoin as like a major, you know, I think he might've predated Steam and all the other guys. I think he was the very first one to do it.
6:07:35Vlad CosteaYeah. I see that there is a Byrne family that's running Overstock. What's his name? Thurston?
6:07:44Karl KrederI thought, I thought it was Preston Byrne is, but the CEO of Overstock.
6:07:52Vlad CosteaIt's Patrick.
6:07:54Karl KrederPatrick. There it is.
6:07:55Vlad CosteaPatrick Byrne.
6:08:00Karl KrederI think he was like the first major like acceptor of Bitcoin.
6:08:09Vlad CosteaYeah. 2013. Yeah. I don't remember it because I was not so involved at the time. Didn't care much, but yeah. Maybe I should get him on my podcast, but he seems kind of important and busy.
6:08:32Karl KrederYeah. He's he's an interesting character. He's politically like active too. So he's definitely like theologically aligned with Bitcoin. That's why he adopted it so early.
6:08:53Vlad CosteaI just wanted to let you know that you flipped Zooko.
6:08:58Karl KrederFair enough.
6:09:00Vlad CosteaSix hours and six minutes. You're at six hours and nine minutes.
6:09:05Karl KrederThere we go. We can wind it down if you want. I think we've covered a lot of ground here. I appreciate you taking the time. Appreciate you listening about Quai. I know some of it can be technical, but I believe it's the nuance of the technical differences that make the things truly work or not. Maybe that's unfortunate given, like you say, you can't send everybody to school and then bring them back in a couple of years. But I think some of the design philosophy comes through even if you don't fully understand the technology of designing so that markets determine resources and they do so in a way that most efficiently allocates the resources to get the greatest amount of scale and decentralization.
6:10:01That's really the design philosophy. We want to be as minimally prescriptive as we possibly can with Quai.
6:10:13Vlad CosteaAnd I sort of respect it because it's trying to do something that wasn't tried for this purpose before, at least not in this way. And this is why open source software exists. This is why we're supposed to experiment with parameters and take advantage of the latest discoveries in computer science and cryptography. I see a lot of smart people in Bitcoin, but it's so desolating to figure out that they will never have their way. Unless there's a major crisis where everyone's going to look for solutions, nobody's going to pay attention to what they're saying.
6:10:59Because the golden goose keeps on laying golden eggs, right? So the number goes up, everyone is happy, everyone's getting rich, so why would they want to change anything?
6:11:11Karl KrederYeah, I mean, I tried engaging in that early on and you quickly realize that's not a fruitful endeavor if you want to make broad changes. If those changes are required to sort of accomplish the ultimate goal, then it's not a good use of our time, you know?
6:11:31Vlad CosteaWell, I know. I don't think there are too many people who interviewed as many frustrated developers as I have. They're like, just this one opcode, I get this one opcode and I'm going to be able to deploy my system. And then they never get the opcode. And it's sad, really, because there is so much exodus of potential and talent that's going to Ethereum or some other chain just because they can do what they want.
6:12:04Karl KrederYeah, right. Well, I mean, you know, I think it's interesting if, I mean, it's almost like sad in some ways. It's like if you could just change like a couple things, it would just be so much closer to being able to do so much more. And then we wouldn't necessarily be in this situation. It's just, it seems intractable to change.
6:12:41Vlad CosteaMy girlfriend is going to kill me because it's past midnight.
6:12:50Karl KrederYeah, well, I'll let you not get killed by your girlfriend. We can wrap it up here.
6:12:57Vlad CosteaFYI, you said you wanted to beat Amir's record. If you really wanted that, you'd have to go for 54 minutes.
6:13:08Karl KrederNo, I'll say it's good here. I don't need to do it for ego or anything. I think we covered most of the ground I wanted to get to. So I'll let you not frustrate your girlfriend.
6:13:26Vlad CosteaMr. Dr. K, where can people follow you and the Quai Network?
6:13:31Karl KrederYeah, so you can follow me on Twitter, handle mechanikalk, spelled K-A-L-K at the end. We also have a pretty active Discord, which you can see by going to our website, which is actually really cool. But our website is qu.ai. There's so many cool things you can do with graphics and browsers, he says. So Ignite the Crypto Revolution and Discord, Twitter. Come join us. Come build if you want to build an application. And you want to know that the chain is going to be able to support your application when it takes off and finds that product market fit. And will do so consistently with low fees. You know, Quai is the place for you.
6:14:33Vlad CosteaDo you have a message for Yonatan from Kaspa, whom I'm going to interview next week?
6:14:42Karl KrederI respect the Kaspa guys a lot. I respect a lot of what Yonatan's done with Kaspa as well as publications. I don't know if we can really have the debate by proxy, but you should ask him why. How would you phrase this in a way that you can translate it? Yeah, very simply, if you want to write the question down, I'll give it to you. Which is, wouldn't Kaspa have faster finalization if you maximized the graph anti-chain?
6:15:45Vlad CosteaThe graph what?
6:15:47Karl KrederThe graph anti-chain. Okay. That's it. That's like the whole question. And if he understands it and then answers it honestly, the answer should be yes. And then if that's true, that means that Quai's graph is better than Kaspa's graph. So you can literally ask one question, you get an answer of him basically saying that Quai's better than Kaspa. That's all I'm saying. At least the graph is.
6:16:13Vlad CosteaSo you wanted to dunk on him.
6:16:15Karl KrederWell, just see where he goes, man. Just see. Right? So very specifically, right? Would Kaspa's consensus be faster if you maximize the graph anti-chain? That's it. The whole question.
6:16:31Vlad CosteaThank you very much, Dr. K. For everyone out there, check out qu.ai if you want to learn more. And I guess that's pretty much it.
6:16:44Karl KrederAwesome. Appreciate you, Vlad.
6:16:46Vlad CosteaThank you all for listening and thanks to the sponsors. Once again, Bitcoin.com News, Citrea, LayerTwo Labs, SideShift, and noones.com. I'll see you on Monday, I think, when I'm going to be interviewing Cory from swan.com, which is an exchange. And he's a very hardcore maximalist. And we're going to have a debate about what this is about. And then on Wednesday, there's going to be the interview with Yonatan, creator of Kaspa. Thank you very much.
6:17:21And keep on listening to good podcasts.
6:17:24Karl KrederThanks, Vlad.