Reinventing Proof-of-Work for Scale with Quai Network

Transcript

GM GM and happy new year. Welcome to what's going to be another fantastic episode of Devantel. So, if you didn't know, Devontel is a 30-minute podcast held every week allowing founders, hackers, anyone in between an opportunity to come on the show and showcase their product. And today, I'm ecstatic to welcome Dr. K, who is the co-founder and CTO of Quain Network. So you don't know qu network is an EVM compatible scalable proof of work at layer 1 that is introducing the hash dollar to the world. So if you stick around for today's episode you'll get to meet Dr. K, learn about qu network and how you can get started today. All right let's get into it. [music] [music]

GMG and happy new year Dr. K. Welcome to the show. ecstatic to have you on, man. >> Yeah, happy new year to you, too. >> Excellent. Excellent. Uh, yeah, thank you so much for uh taking uh the time out of your your day uh to come chat with us. Uh really excited to learn about Quay Network. I know a lot of folks tuning in today are as well. Um but uh I guess before we get into all all that good stuff you have prepared for us, uh would you like to give a brief introduction about yourself? >> Uh sure. Yeah, my name's uh Dr. K. um been in uh the crypto space at least uh as a hobbyist since about 2012. I've been working full-time since 2017.

Um kind of got my full-time start at Consensus called Grid Plus designed and built the lattice one hardware wallet. Um sometime in 2018 came up with this concept of how to scale proof of work blockchain. So um I managed to actually get a National Science Foundation grant uh to work on that as sort of like a research project that was uh after I graduated my doctorate in engineering. So it was kind of like a posttock uh project kind of thing. Um and uh that eventually turned into a full-time thing. Um so um been working on um Quai full-time since 22. Um and it took us from concept in 2018 to

implementation or mainet took about seven years. Uh full-time it was about three years to sort of implement um quai. Uh yeah so that's kind of my background at least uh in crypto. Um yeah. >> Awesome. Awesome. And yeah uh uh I I guess what was the the compelling thing um that kind of brought you into the world of crypto? like what was the thing the like oh wow this is like really cool kind of moment that yeah like makes you want to like be here full time >> yeah so from my perspective right um Satoshi really just summed it up in the title of the white paper peerto-peer electronic cash right um and the sort of libertarian underpinnings of that and

obviously um relating that with proof of work meaning you have like a physics-based money or an energy based money in some way. Um certainly answers a lot of the questions related to rampid sort of monetary policy and political corruption. So, you know, all of those sort of ideas about, you know, money, politics, and society, uh, and individual freedom is what drew me into, um, the space in terms of kind of where I saw it go after really like 2017 is um, you know, it's a very technically challenging problem, but I don't think the industry was going down a path of fulfilling that sort of promise of

peer-to-peer electronic cash. if it wasn't, you know, privacy not being there or scale not being there while maintaining some form of decentralization. Um so that ultimately led me to want to develop Quai which is um you know scalable proofof work system. So sort of maintain um the really good characteristics of Bitcoin um all proof of work relating uh the minting cost of a token to actual energetic input costs uh as well as having a relatively decentralized system but taking that to scale so you can actually use it as peer-to-peer electronic cash and then obviously to do that you also need some privacy features but but that is you know wanting to

actually see that vision brought to true life is is really what motivated me to start Quai. >> Awesome. And uh yeah, you you didn't found uh Qui Network alone, right? Uh there's um a couple co-founders if I'm not mistaken that uh uh came along for the journey with you and uh um in particular at the University of uh Austin as you mentioned. I find that really interesting like how um such a such a big network kind of started out of a university. That's not usually the traditional uh startup route, right? Uh would you want to get into that just a little bit like to hear the background story there?

>> Uh yeah, so I got my my engineering degree at UT. Um so I finished in like early 17. Um my doctorate and um I kind of came up with this idea how to scale proof of work in 18 and I shared it with one of the professors that I knew at UT. Um so he kind of had a research group in electrical engineering um you know looking at distributed systems. So he's like yeah let's let's uh you know start looking at this. So he had a couple grad students, a couple undergrad students um and we kind of started working on it and we got that National Science Foundation grant and uh yeah and then

eventually in 22 we managed to raise money uh from Polychain Capital which allowed us to really go like full-time into it with a much bigger team. So that's kind of kind of how it started. Yeah. >> Cool. Cool. Yeah, that's amazing. Uh and yeah, I guess uh no better time than now to kind of get into the details of this. Um so uh one of the big things I've seen um reading reading up on Quay Network is this notion of proof proof of entropy minima poem for short. Um can you just explain what that is and um why you think uh why you guys think this is better than traditional uh proof of work?

Yeah. So to get to poem, um I just need to give a little bit of a background. Um so Quai is a hierarchy of merge minds blockchains. So essentially what we do is um if you're familiar with like a super block um it's kind of like this concept when you're mining a block, you can hit a difficulty, right? If you hit the difficulty, you find a block. Um, but then you can have a concept of a super block which is instead of just meeting the block difficulty, what if I meet the block difficulty plus like five extra zeros, right? Then what we can do with that is we can make like a super chain. So you kind of have like this lower chain, then you have like an upper chain, you have a chain above that. We call that prime region and zone. And then what that allows us to do is it

allows us actually to create a pyramid of chains where prime is at the top, then you have regions below that, you have zones below that. Um now then what ends up happening is as a minor you have to sort of operate a slice meaning anytime you propose a header you have to propose a prime region and zone sort of composite header and then depending on what difficulty you get you'll either get a zone block a zone region block or a zone region prime block and that's how you have sort of like multiple threads but they're all working on a common consensus layer which is just proof of work and and what poem does is it allows a deterministic outcome without mult multiple rounds of communication. So in traditional proof of work, you can get a fork, right? Um and basically what

happens is it's a natural thing in the system because when once you find a block, it takes time to propagate throughout that system. Uh during that propagation time, someone else may find a block just like a very super honest orphan as we might think of it in Bitcoin or an uncle in Ethereum. Um, and so the problem with that though is that you'll eventually get resolution, but it takes time. And so as we add more of these threads, so to speak, more of these shards, the ability to wait to get like the canonical tip limits the total number of shards you can have. Because if I'm in one shard and like statistically maybe it takes me sort of like 30 seconds to kind of know

that we're all in agreement and then I have another shard and it takes 30 seconds to know that we have agreement. Now when I start trying to go back and forth and this aggregate I have this like aggregate composite agreement time which may be like a couple minutes right and then as I keep adding shards like it just gets longer and longer and longer in time. So what you really want to do is you want to make sure that your consensus mechanism is deterministic. So we don't have to have rounds of communication and sort of statistical convergence. We want to converge as quickly as physically possible. So basically what poem does instead of just measuring or rewarding blocks um you know the difficulty threshold it uses um the whole hash basically and we

basically take the field size minus the whole hash and we take a log of that and then we call that bits. So you know a bitcoin block has like 60 leading zeros. So that would be like 60 bits, but depending on sort of how lucky you get when you find a block, you'll get some fractional bits. So it'll be like 61.25684, you know, whatever to whatever precision you took your logarithm, right? Or you might even get lucky and you'll find 64.89257 bits. But the interesting part here is because you're maintaining um the the um sort of randomness of the field, you basically have a guarantee that everyone will choose the same winner um to the

point of a hash collision. So like if two blocks are mined, unless they like literally collide hash, there will be one better block than the other. and the betterness or the weightiness of that block adds up over the sequence if that makes sense. So what poem does is it allows uh a deterministic fork choice rule that is also compatible uh with extending not just one but many chains simultaneously. >> Wow. Yeah, that's amazing. And um that I I I suppose that took quite a bit of um testing uh on a test net and a lot of uh a lot of back and forth and revisions until it got right. Right. And

>> yeah well so so what what ended up happening to be honest is you know so so Quai kind of like evolved over over time. Right. Um the very first thing that like started Quai was the concept of just the hierarchy merge my chains and we kind of like built that and then when we got there we hadn't really thought too deeply about consensus. We had come up with something that we called the hierarchical longest chain rule which was basically saying like we'll pick the heaviest prime then heaviest region then like heaviest zone sequentially. But that actually creates like a um like an unresolvable loop if you have many chains that are propagating in time. So you can get these situations where you're sort of like trying to compete with the chains

and you need to resolve who the leader is because in the system like prime is not the leader. Prime is the follower, right? The leaders are simultaneously all of the zone chains, right? So it's it's a it's a multi-leer system and the nature of that multi-leer system required us to figure out a way to create a deterministic choice rule that would be compatible with having many leaders, right? And that's kind of where POE came from. So we actually built the system and spent like a couple years putting it together before we realized we needed something like poem and we were just lucky enough that we could figure something out that actually kind of worked. Um there there's there's sort of like an interesting like knife's edge that we're kind of balancing on and that

you know depending on like how academic or you know how theoretically minded you get um there's this concept of like synchrony versus asynchrony. It would be our sort of um assertion that we're very much on the edge of asynchrony here because you you have a many liter system and then this rule will always resolve out once everybody sees the data. So given consistency we can guarantee consensus and that like technically should barely exist within sort of the theoretical frameworks that have been proven within information systems. So, we kind of got lucky, I guess, would be [laughter] the statement.

>> Hey, man. So, sometimes you just need to get lucky that one time, right? So, [laughter] um and and yeah, I I guess uh a couple couple things before we uh move on. So, um just curious like what um like mo most of the at least on the Ethereum side of things, mo most of the interest industry has moved on from proof of work to um proof of stake. like what was um what was the determining factor from your guys' end where you're like no we don't think proof of stake is is the future um and you guys wanted to iterate on proof of work. [snorts] >> Yeah. So there's kind of like um there's there's kind of like um

a very a visceral sort of divide between proof of work and proof of stake. Um, and it has to do with the fact that proof of work at least instantaneously is hard money, right? So like in a proof of work system, so like let's talk about gold, right? What gives gold value? The thing that fundamentally gives gold value is the energy that goes into finding, mining, and refining it. So over time sort of the market value of gold is the cost of producing gold. Um in proof of work at least instantaneously that is true. Uh so like if you look at bitcoin right the cost of

making a bitcoin is the cost of a bitcoin instantaneously like that's kind of it trends to that. So that property of hard money um you know is very much part of it right and then the property of immutability meaning um economic immutability right so something just becomes inordinately expensive with time to ever go back and change versus you know having a committee being able to effectively resign some things to change history uh which doesn't have a high economic cost. um that's kind of initially what you know kept us focused on proof of work. But what I will say is the nature of poem and the nature that it is uniquely derable from work and

really nothing else. Um actually makes proof of work much more scalable and much faster. um I cannot find an analog in proof of stake that would allow a many liter system like we can do with poem. So you can't even scale proof of stake to the same degree that you can scale proof of work. Um and one other way to think about that is with poem um we never have to talk about which block to pick. We just need to broadcast the blocks to each other. With proof of stake, we have to talk about it, right?

So like I have to get a proposal, I have to sign it, then I have to send it to you, then you have to sign it, and then we have to aggregate the signatures and be like, "Okay, quorum. Now we're good." With proof of work, it's a broadcast only system. So like once the the block hits everybody's nodes, like because the rule is deterministic, we all know that we're in agreement. So there's no need to talk. We just need to make sure that we see all the blocks, right? So it's a much much faster consensus mechanism. It's a consensus mechanism that allows the shard sharding because it's deterministic. You you can't recreate that in proof of stake. And if you try to, you actually get a system that turns back into mining because I've thought about this actually quite a bit. So if you tried to use signatures to do this,

you would actually get people grinding signatures. Um so it it would devolve right back into mining. >> Fascinating. Fascinating. >> Yeah. And uh and yeah, I guess you you kind of touched on it as you were explaining, but just the scalability factor. Um I was reading uh up on your doc up on your guys' documentation that um uh this uh poem poem consensus mechanism kind of um solves the blockchain trialma, right, of being able to kind of scale scale a blockchain. Uh so uh is that is that true?

>> Yeah. So, so basically what happens is within that hierarchy, you kind of have different um parameters so to speak in like your sort of trilmma, but because it's all in one system, you get like the best of everything because you can put points in multiple parts of the triangle. Um if if that kind of makes sense. Um so what what ends up allowing us to scale, right, is we have like a zone chain and we're operating those at 5-second blocks. Um and so you can get you know Visa Mastercard like uh settlement times in a in one of these chains. Um and then you know one of those chains we can do around 700. We can probably go more than that per

chain. Um but let's just say 700 for right now. Now then what what's interesting is because of poem and because of this hierarchy once you have like a chain fully saturated you just kind of add another chain and they coordinate via the next sort of what we call a dominant chain. So a region or prime and so you can just keep the system expanding and all these chains are still interoperable with each other because they're all still running the region and prime chains together. Now the sacrifice here is that if I want to do a cross-chain transaction in these zones, there's some clearing time for that. But if we went to say like the whole western world is using this thing, sort of your worst place, well I I wouldn't call confirmation time. I'd

call it settlement time, which is different, is about 90 minutes. So if you're sort of like in the worst case farthest zone away from me in the world and I try to send you money in this um situation where everybody with internet is like using Kawaii as a means of payment um you would in five seconds know that I paid you money but it might take 90 minutes for you to get that money in a way that you can respend it if that makes sense. So that's like the sacrifice. But that again is if you're doing a 100,000 transactions per second constantly and you're at this like 100 chart topology, you might have to wait 90 minutes to be able to respend your

money. You're still going to get that 5-second payment guarantee, but you'll take 90 minutes to spend it. Now compare that to today when you actually like pay a merchant with a credit card. It can take them 90 days to see that money, [laughter] right? Um, so you you know I I don't think um like that 90minut sacrifice is actually a sacrifice at all, right? As long as you're getting an economic sort of confirmation within 5 seconds like you're good. It's just um waiting 90 minutes to be able to respend. There's very very few economic use cases that need that high velocity of money. So we don't even really see it as a compromise to be honest.

>> Yeah, that doesn't sound like a compromise to me. I mean it's better than a wire transfer. Um and and yeah, like you said, 90 90 minutes versus 90 days. It's 90 days. >> Well, and and and the other thing too is what what we see is that the way the system is set up is it actually encourages um geographic uh subnetting relative to the shards uh from both a node and minor perspective as well as a transactor perspective. So, what you might see is that, you know, as the system grows, maybe the east coast is in one chart and the west coast is in another, right? So, then when you're in New York and you're buying your bagel because you're from New York, um, not only are you going to get a confirmation in 5 seconds, the merchant's going to

receive their money in 5 seconds. It's only when you go to like the West Coast, it might take 20 seconds. And then when you go to China, it might take like a minute for the merchant to actually get the money, but they'll still know that they're getting the money within 5 seconds, right? So that that like 90minute problem is only going to be like the absolute worst case when you're kind of in a very unoptimized u situation, but the incentive of both the users and the node runners is to aggregate together, right? So in most cases, even with the system scales, 95% of the time, you're still not only going to be getting confirmations in five seconds, you're gonna get settlement in five seconds because you're going to be in a common shard.

>> That's amazing. That's that's pretty nifty. Not going to lie. [laughter] Um and uh I I guess shifting gears just a bit um to the the development side of things. Um it uh it seems like um Quiet Network is an EVM compatible uh chain as well. So um for the EVM developers watching today, how easy is it u for them to take their existing smart contracts and deploy it on quiet network? Is there is there any um anything that the EVM doesn't support on uh quiet network? Um so the only difference is that we have a sharded address space. Um so we have uh

variations of hard hat and VM that like allow people to sort of port their contracts over. Um they just have to deal with the fact that we're in a sharded address space. So the way that sort of this hierarchy works and and this kind of gets into a meta these technically aren't different chains. uh they're all actually sharing a common state route. Um and it's divided by the first um two byte prefix, right? So a 000 is sort of the first shard. A 01 address is the second shard, a 02, you know, so on and so forth. Um but but that means you you actually technically have like one state route that you're updating across like the whole system

even though it's technically sharded. It's just which chain has permission to update what part of the state route, right? Um so so that's kind of the complexity that that is added. It's just the um as a developer you have to use kind of like our tooling because we deal you sharded addresses um and correctly. Now um right now right this this system expands dynamically. So currently there's only one shard in the system because we're not at 700 DPS yet. um the system dynamically understands its uncle rate and as the uncle rate increases it allows another shard to take place. So basically the uncle rate is like a it's an indicator that this shard is full and

we need another shard. So the system automatically adds shards as the uncle rate gets above a certain threshold for like a period of time. So then it just expands. Now all of the um like cross shard interactions you can do them uh between contracts but you have to specify like what your sister contract is. So right now deploying in a single shard setting it looks pretty much like a copy paste for EVM like contracts. Um if you want to support a multi-shard setting in anticipation then when when we shard um you can do that as well. you just have to write basically create like a list of sister contracts when you deploy this contract and then when the

new shard gets deployed you have to deploy the sister contract. Um but but that's that's about it. So like for for 90% of people unless they're like very forwardlooking they don't need to worry about the sharding at all. Um for the other people they just have to use kind of the tooling that we have to deploy in the first shard. >> Gotcha. Gotcha. And yeah, uh uh to this point, uh my follow-up question was um whether or not there are any available um SDKs or specialized tools for um people to pick up and and uh play around with Coin Network. >> Yeah. So, right now we have uh what we call quis. It's a fork of ethers. Um it's a fork of it was originally fork of ethers 6 um v6.

Um we've modified it a bit obviously for for our purposes. Um, we have like a much better subscription model. You like honestly like the [laughter] the way Ethereum uses endpoints is utter madness. Um, we think we've done a little bit of a better job. So you you actually have instead of like a polling model, right? So if you use uh ethers or you use any of like the SDKs, um, they don't actually use subscriptions. It might look like you're subscribing to something, but like behind the the SDK layer, they're actually running a pulling on the node.

So like if you're like, I want to know my balance, you're just asking a node for your balance every 10 seconds, which is like wild in terms of the amount of overhead that induces on the nodes. We actually have like subscriptions. So like you subscribe to addresses or you subscribe to contracts and then the node like will update you and it will leave that connection lively you know for you know 30 minutes or whatever and then you'll have to refresh the connection but you only have to like refresh it on the timeout. You don't sit here like pulling all the time. So you can lower your your your request by like a thousand or more 10 like 10,000x right.

Um but but otherwise from the user's perspective it looks pretty much identical to ethers. just has our own sort of like name space. >> Right on. Right on. Uh good stuff. And um I guess in terms of um like the uh the long-term vision of of of Quiet Network, is there anything that you can can share um on the the road map um before looking into this here? Um yeah. So I mean basically the interesting thing about why is right we can do stuff in we can build applications with web 3 that look like web 2 but we're doing it

on a proof work system and that proof of work system is scalable. So from a developers perspective the reason that quai is very interesting is because if you look at Ethereum when it gets sort of a hit application um fees precipitously rise and that sort of hamstrings that application's growth um so much so that a lot of Ethereum's um apps have started to go onto app chains right you know unis swap being a great example um you know there's there's plenty of them but but maybe that's the most recent one Um, right. The thing with Quai is because we can shard, because we can scale, if you're a developer and you build an app, the network grows with

your app. And so you're kind of guaranteed this sort of one cent per transaction availability today, tomorrow, and in five years, right? So that's sort of like a super powerful thing. And we've been building applications on Kawaii that take advantage of that. Um, one such application we have is called Kipper. Um, it's pretty cool. It's like a tipping app on X. Um, you can use it on desktop. There's also um there's there you can use it on um Android. There's an app there's a beta on on iOS.

We're having issues getting it through their store, but it's really cool. You basically can go if you like you can set it when you like a tweet you actually like send somebody one quai or you can like tip them and you know write them a message and then your reply actually shows up at the top of their feed um kind of like highlighted in things. So you know but but that's just an example of what you can do when you have that like very guaranteed sort of low fee high throughput system. Um, another thing that we've done, um, which is very interesting economically with this system is called SOAP, um, a subsidy open market acquisition protocol. Um, basically what that is is we're merge mining um, with Litecoin, Dogecoin, uh,

Bitcoin Cash, and Ravencoin now. Um and so basically what ends up happening though which makes this different from merge mining is um we only pay the minor Quai and the merge mine token comes back to sort of like a foundation address. That address then liquidates the secondary token to buy quiet at market. So what that does is it actually creates much better uh proofof work economics for us because you don't have constant minor sell pressure. But the reason that we can do that is because we're in this very high-speed, high throughput proofof work system and all these other proof of work systems are much much slower. So we can kind of take advantage of what we like to think of as underutilized hash

rate to increase our security for free. Um right so we basically have like a net neutral emissions at this point, right? Because if we pay out $100 worth of quiet, we're getting a $100 of another token back. So like on a net net basis, we don't even pay for our security. Like it's free for us. um like we might be paying like 2% or something, but it's it's it's like a tiny amount, right? Um we're also launching something soon that's kind of an extension of that where Qui just becomes like a pool, but it can become a pool for any chain that exists. Um so those are like more examples of applications that you can just build because Qui is high speed and scalable. Um yeah.

Yeah, that yeah, I think Quai Network is Yeah, it sounds like quite uh quite the beast of a network and um yeah, like you mentioned there there's lots of opportunities for people to to build um on it. Um I guess from that perspective um are there any grant opportunities or or um upcoming hackathons or anything like that to make it um uh to to entice uh developers to to get on the the Quiet Network train? Yeah. So, we have uh we have grants. Uh we have retroactive grants. So, if you go to our docs page, you can find out all about our grants.

Um yeah, I mean, we have, you know, people have built NFT marketplaces. They've built um you know, like unis swap type things. They've built pumpf fun type things. Um the the there's a couple that we're we're more excited about too. Um and if someone wants to work on it, um that'd be amazing. Uh BA basically one of the the implications of SOAP which is very interesting is we end up getting loose referencing to the blockchains which we merge mine with. So that actually creates a mechanism by which you could trustlessly wrap any of those tokens in and out of Quai um without

needing like a custodial intermediary. So basically like Bitcoin can sort of be wrapped into Quai, Bitcoin Cash can be wrapped into Quai, Dogecoin can be wrapped into Quai, right? So if somebody kind of wants to build that service where you could go back and forth between those assets, that'd be super interesting. If someone doesn't do it, we're going to end up doing it because what we'd like to see is um the SOAP revenue we get sort of guarantees volume every day, right? We're doing like 20,000 or so in volume every day um in proof of work. So, right now we're having to clear that volume through centralized exchanges, but if we can have that wrapping service, we'd like to clear all of that volume through a DEX that runs on Quai. Um, would would be

pretty cool. And then the cool part is that volume goes up with our hash rate. And as our hash rate goes up, we get more subsidy, which buys back more token, right? So then our price goes up. And when the price goes up, the hash rate goes up, which gets us more subsidy. So, it's this cool sort of virtuous loop of sort of increasing hash and price, which hasn't like ever existed in proof of work. It's always been the opposite, which is um you know, miners mine, they have to sell to cover costs. Um and it like causes the token price to go down unless you can get more new buyers than you're having to sell to cover your security every day. So, like we're really excited about soap because it really changes the the economic game.

I think it actually makes proof of work um more economically like robust than proof of stake at that point because even within proof of stake you still got to pay 3 or 4% to your stakers. In this case, we're getting security but like we're net neutral which which is pretty cool. >> Yeah. Yeah. That that that is pretty cool. And uh yeah, for anybody who's watching today or or listening um there's no of opportunity on coin network and I mean building a dex that's a money printing machine. So [laughter] >> build building a dex with guaranteed flow.

>> Exactly. >> So like right now we're doing $20,000 a day between those pairs, right? >> But you know if our price goes up again we'll be doing $40,000 a day guaranteed every day that like you got to clear, right? So, um, you know, there's there's there's a lot of opportunity there. >> Indeed. Indeed, there is. And I guess, uh, um, just before we let you go here, um, what's the best way for folks to kind of, um, keep up to date with all the latest and greatest from, uh, Qui Network and and yourself? >> Yeah. Um, so they can follow Quai. It's a Quai Network on X. Uh, they can also follow me. Um, my handle is mechanical K spelled K- A K at the end. Um, if you

want to learn about the grants program or you want to learn more about Quai, just go to docs.qu.ai. Um, you'll find all the resources you need to run node, develop an app, mine, um, get a grant. So, >> amazing. And yeah, we'll have all those resources for you in the description below. Um, so definitely check that out. Um, get involved in choir network. check out the docs, start building things, move over some smart contracts from the EVM world here. And uh yeah, I'm really excited to see uh Qu Network continue to to uh grow here. And uh thank you again, Dr. K, so much for taking the time out of your busy day to come chat with us today.

>> Appreciate it. >> Of course. Of course. Uh with that, I wish everybody a uh very happy new year again. Uh happy weekend wherever you may be. Uh, and we will catch you back here for another episode of Devon next week. Till then, have a good one, folks. >> Appreciate it. >> Of course. No problem.