Debate PoW vs PoS with Dr. K. (Quai Network) and Justin Bons (Cyber Capital) at Common S3nse - 2026

[music] Now, um, last but not the least, we're going to have a very, very interesting debate. Just to, um, refresh you on what we're going to talk about, we're going to have a debate about proof of work versus proof of stake. Even I cannot wait to hear what is going to go on here and and to um moderate this very interesting um debate. I'm going to welcome Joel again once to the stage and um yeah,

thank you.

Thank you everyone. Um this is a fun eternal question and this should be a nice spicy way to round out your day. Um I'd like to call our two contestants for this here game show up to their positions. Um, in the interest of everyone, try to be respectful and semi- take turns. I'll try to direct the conversation a little bit, but mostly let there be a discussion and debate.

All right. So, first of all, um, I would like you gentlemen, even though everyone should know who you are, to just introduce yourselves first. Let's go with you, Dr. K.

Um, my name is Dr. K. I'm the co-founder and uh CTO of Quai Network. Um I don't uh usually identify as a maximalist in many things but uh over the years I guess uh I have become a proofof work maximalist. Um and so I will be representing uh that side of the argument.

[snorts]

Give this man an applause.

Yay.

Thank you. really like out of anyone that I could debate proof of stake versus proof of work, this is the best champion right here. So, uh here's my respect big time. So, I'm I'm Justin Bonds. I'm the founder of Cyber Capital, which is Europe's oldest uh liquid token fund.

Um and I'm a researcher and investor. And uh I like to have debates as you all know. So, this is going to be a really fun one. I'm actually uh hope you all enjoy it.

Yes. So, let's start out by just briefly stating your position, but also I'm sure there's going to be proof of work is better, proof of stake is better. Also define what you mean by better. Like what does that mean? What's the criteria for that?

Like more economically efficient, more decentralized, more censorship resistant. Like what does it mean better so we're on the same page of what you're fighting over? So, who wants to go first? I'll go. So I will argue that proof of stake is in fact more secure, more decentralized and more economically sound.

And I will concede on a lot of other points here, I'm sure, when it comes to proof of work. But I will argue that these three attributes are the ones that we should prioritize. And that is why I champion proof of stake as the solution that we need as a movement. Okay. Um I I would uh say that uh proof of work is better uh because it's more scalable.

Um it has economic incentives that in the long run uh make it more uh decentralized. um and that it is a system which is most resilient to um nation state level challenges. Um so when when I think about crypto systems fundamentally sort of the goal that and the reason I got interested in crypto was sort of the libertarian ideals of creating parallel money systems uh which means you need some scale obviously you need some privacy um but you can't compromise on the decentralization and if we're truly to succeed creating digital cash we have to understand who our adversary is in the mid mid to long term which I always posit as a nation state. So I asked the question in the design of these systems in the design of consensus what mechanism potentially can stand up to that type of attack given the economic incentives of the system. So there's there's a little bit longer but

in principle I agree with everything you said there in terms of like what what do we define as good? What should the system achieve? I think the difference is is you think proof of work achieves this. I think proof of stake achieves this better clearly.

Um yeah I mean I I guess I guess that that is the uh the difference. Um

so there's a lot of ways we can approach this now right a lot of different arguments. Um I suppose I'll go with the economic argument. I think that is one of the stronger ones I think for proof of stake. uh if you so in this example I'll say if you take an equivalent network uh of in terms of market capitalization inflation and fees uh the proof ofstake network is orders of magnitude more expensive to attack and I'll use a term here the cost to attack a proof of stake uh network of an equivalent size is far far higher so like that is something I value from a security perspective and also specifically From the perspective, as you say, potential nation states or other bad actors attacking the chain, putting a dollar value to how much it will cost to attack this blockchain is how I believe these blockchains are actually protected. These are incentive mechanisms.

After all,

um I want to start with a little bit of history and it will play into the sort of response the the economic argument. Um, if we look at money and the history of money, we had gold and gold was the Neanderthal version of proof of work. The reason a gold coin has value is because over time and space because of how gold was distributed throughout the earth's crust, you needed roughly the same amount of energy to go find, mine, and refine it. So presenting somebody one ounce of gold 200 years ago or today is a representation of work that went into that asset. That is why Bitcoin uh gets the moniker digital gold.

And potentially, you know, I don't fully agree with it as digital gold. It's a little bit more complicated than that, but instantaneously the production of a coin given work where the value of that coin is roughly equivalent to the cost to produce it is a goldike um property. Now if we look at history, what happened to gold? Gold had some problems with it. It was hard to store, transport and divide.

So that gave opportunity to create a marginally more um uh easy to use system which was notes, right? And then once you had gold reserved by the banks and notes, the bankers eventually figured out how to fractionally reserve that. And then eventually we moved off that system entirely because they failed. So often to such a degree of fractional reserve, we went into fiat. But fundamentally in the greatest age of America for a hundred-year time period when we worked off the gold silver biometallic standard that was a proofof workbased system and over the following hundred years we incrementally worked our way into a federal reservebased system of fiat currency which is a set of commercial banks who are members of the Federal Reserve who get to mint the tokens that have no backing.

that circulate. So you have the people that have capital, the stakers determining the rules of the system and who gets allocation of capital and the money supply. The whole point that we are where we are today is because of that economic set of incentives and outcomes that happens in a stake-based world. The reason that gold was so interesting and kept everybody honest is because it was tied back to the physics of work. So that's a manifestation in this historical allegory which I think is fundamentally different between work and stake.

Now to address the point of cost, the issue of cost is not necessarily a direct comparison between proof of work and proof of stake. I would argue that what we're really looking at is how much would it cost to operate a technologically sufficient proofof work system that can work as money globally relative to the current global money system. So if we do some reasonable projections of that, you might come out with you need 1% of electricity going into a proofof work mechanism, but that's replacing potentially 20% of GDP that's taken from us through fees, through taxation, and through inflation.

A lot of the things you're saying here are like nice and philosophically very beautiful, but they have often little bearing on the argument, right? Like I I I did hear the piece about you think that that the current system is a stakebased system etc. and that proof of work represents this revolution etc. Right? I understand that that is a argument but all this all this other beautiful speech because we're having a debate here I agree with like it's it it doesn't have any actual bearing on this argument.

Let me let me

finish the point then. Your argument about the cost of the system I don't think matters because if we can

the cost to attack the system doesn't matter.

Sorry. The cost to administer the system. The cost to attack the system is a different argument. So

I'm making the cost to attack the system argument here.

Not the cost to operate the system.

Correct.

Okay. So

now, isn't it fair that both of those arguments apply to this overall argument that because I know Justin, I've heard you say in the past that uh proof of work is more inefficient. It's more expensive than it needs to be. That sounds like not just a cost of security, but also an administrative cost, right?

Yeah, I think that's fair. Yeah, it's fair. But but I think I think in here like like I I find a lot of these philosophical ideals appealing. But at the end of the day, and this is what I'm trying to do here to kind of bring this back to some common sense, it's like if the system is more expensive to attack, it is more secure. Like if you say, "Oh, it's work is not inherently more secure than stake."

You can from the attacker's perspective, you just put a dollar value to it. And when you put a dollar value to it, proof of stake is by several orders of magnitude more expensive to attack and therefore more secure. Okay. So, I think that there's there's a nuanced argument around security and what security means in long-term equilibrium. So, so I I will agree this is a there's many aspects to security and this is one aspect of security which is economic security.

Right.

Fair. So the ultimate security that I think you need to get to um is actually bootstrapped by the consensus mechanism. It isn't actually the consensus mechanism long term. And the reason I say that is when we talk about decentralization, we're talking about this concept of the cost to attack, but that cost to attack takes many different forms.

Well, I would prefer if we separate concepts here a little bit. Could we separate the cost to attack from the decentralization of the system? I would find that a lot more productive for this conversation.

Well, I I think that they're intertwined. They are they are intertwined but they they can also be measured separately and you can have radically uh uh different uh variations on that theme right you can have a system that is very decentralized with a very low cost to attack and vice versa

I think that it's um it's a continuum of two variables right so if we think of it as like x and y

can we define decentralization I would define decentralization as the distribution of power Um and if we measure the distribution of power that that will be an economic component right

one of the components like I tend to separate this like personally uh so like I would look at this is the economic cast cost to attack the blockchain and then I would have my decentralization measurement and analysis as like a separate me metric. I mean yes they're interrelated and the uh the economics absolutely drive a lot of the incentives and I absolutely do believe that. Um however I think there is a difference like but like like you could have massive economic security and you could have two nodes right with an admin key like like you know we have to separate these things a little bit.

Sure. Sure. Let's um let's foliate this but maybe we we can circle back on the the long tale of security. Um but I'll I'll directly address the cost to attack. Um so when we look at sort of proof of work versus proof of stake um there's this sort of uh continuum in work and it's not just are you proof of work or are you proof of stake on the very extreme of that continuum you have something like random X and random X is a CPUbased algorithm.

If you can get a botnet or you can rent compute, you can just take orders of magnitude more compute that's in the system, pay for it on a very temporary basis without having to invest in any capital and you can attack the system.

So it's more apples to apples for cost to attack. It's like buying coins versus buying hash, right? It's very linear.

Yeah. Yeah. It it is that it is linear, but but then as as we go along the continuum, you get down to more specialized hardware. So then you get down to GPUs. GPUs are pretty prolific, so they're a lot like CPUs nowadays.

But then you get into AS6. And AS6 are an interesting hybrid model because I would argue that they're staky in nature. Meaning that for you to participate in the system, you're taking investment on capital that has a 12 to 18month time frame to get return on your capital at least. So you're coming in, you're buying this, you know, $40,000 ASIC. You have expectation to get return.

In some ways, that is stake. And that plays into the the actual cost.

I will 100% agree with you. That is stake. And my whole point is the stake for proof of work system is lower far lower. And that's exactly my point because I am drawing an equivalence between these two things.

Fair. So, so the the the point that I make though is that when you are in a proof of work setting, you you have the hash rate and you have how that hash rate is produced creating a stake based mechanism that gives you some amount of security budget and as long as that security budget is impractical to attack over a midtime frame before this is ubiquously adopted as money because once that happens I think you're in a completely different paradigm with a completely different mechanism for security which is actually social in nature because if people use these systems as money, it becomes politically untenable to attack it if a mother's buying her groceries and filling up her gas tank. So in the longest term of these systems as well,

well, but the difference is and this is the primary difference is if we can end up at the same end point and we're saying that's true. The problem is proof of stake pulls us back into the economic incentives that reflect what we currently have today.

I'm going to push back. I'm going to I'm going to push back. Okay. I'm going to push back on that one very hard because I feel that is a I want to also discuss decentralization separately as well. But I want to push back because you you mentioned this idea earlier.

It's like this framing that oh proof of work is the cipher punk revolution and proof of stake is just the repetition of the old system. Nothing could be further from the truth when it comes down to it. The old system is based on power and the monopoly of violence and democracies which are one person one vote right like that. Those are the systems we live under today. It is not an equitybased system.

When I'm talking about stakeholder governance I'm talking about one coin one vote which is also how stocks work and private companies work. It's a very good and functional pre precedent for how to do governance and how to coordinate and organize. I would argue if you look at the current financial system, the incentives that take place in non-competitive accretions of capital through interest create centralized system where power is concentrated and more importantly it allows them to change the rules of the system arbitrarily against everybody else that doesn't have the majority of the voice in that system. That

that is the biggest point.

Sure. I'm going to I I'll concede. But um what I'll do there, right, is is um let's that's the perfect bridge to decentralization, right? Because the reality of proof of work today and when we look at proof of work systems at scale, who controls a proof of work system, it's a few industrialists in at scale. That's how proof of work works.

That's what we see in Bitcoin. like a normal person on mass don't have any say because in proof of work there's economies of scale right in proof of stake there are no economies of scale. So in proof of stake what that means is everyone can participate in the system no matter how much capital you have. You can you can have a say right and and that say collectively can actually be significant. But due to the nature of proof of work and the economies of scale if you play that out over the long run what does proof of work look like?

a couple of industrialists, elitist, a couple of massive mining companies, they control the majority of hash power like like like ultimately that's the distribution that proof of work heads towards and if you contrast that to something like proof of stake like just look at Ethereum you know there's over 5,000 uh validators right and and like or you know the Nakamoto coefficient for proof of work systems tend to be extensively low and I will concede Quay is one of the best proof of work systems I've ever seen. So I'm sure you can bring some exceptions to what I've said here, but but I think the principle like holds true and it's fundamentally because of the difference of what these systems are. In the case of proof of work, right, all of the security, all of the incentives are derived from the inflation and fees of the system, right? In a proof ofstake system, all of the incentives, all of the security are derived from the inflation and fees of the system plus the capital of the cryptocurrency. itself.

That's what makes it civil of civil magnitudes more secure, right? And and and the idea that like stakeholders get to have a say like if you own part of a network, then you should be the one that has a say over that network, right? This idea that that that a we have a completely separate ruling class which we call miners is not a particularly democratic concept at all especially when you factor in the reality of economies of scale and and the effect that has in terms of the centralization of the mining industry which we've seen time and time again and ASEX even worsened that equation with the limited amount of foundaries there are in the world. So I think if you look at the centralization surrounding Bitcoin mining, it's it's much more subtle than uh you're making it out to be. However, it is an outcome that exists in what I would class as a technologically insufficient proofof work implementation.

So the fact of the matter is you have 10-minute block times, right? You have one type of hashing algorithm. um you have rewards that mean everyone necessarily must participate in the pool if they want to be able to mine at all at any you know non

quai fixes that issue as far as I'm aware as well right

correct so qui basically has much faster blocks we also inter do something called a work share which is like a pool share so you effectively have a pool in chain um we have a system where every node in the system itself can actually be paid to be a proposer for anybody that wants to mine. So your barrier to mining doesn't even have to be running a node. And even if you're running a single GPU, you're still going to get a payout every day or so without having to participate in a pool. So that degree of centralization around the pools uh doesn't exist in a technologically sufficient implementation. And then in terms of the diversity, the workshares actually allow us to go multi-algo.

So we can do u GPU mining, we can do script mining, we can also do shaw mining. And that hybridization causes you to get better diversification over the producers and the foundaries so that you're not depending on on any single source. So I I think in the right implementations the um

but none of this fixes the economies of scale problem because the economies of scale like you cannot run a GPU as efficiently as a data center can

as a normal user.

Well well so so the

you can stake your coins and have the same benefit as the elites. No mining is fundamentally an eliter system if you frame it in this way.

Hold on, hold on. This this is this is actually not true for Qui. So the bigger that Qui gets and the more shards that it has to maintain, miners actually have a incentive to topologically self-organize into regional subgroups. And in doing so, they're actually going to increase their um expected revenue per hash. So you can't have a centralized mining system in West Texas running 80% of the network hash rate over multiple shards.

You just put a data center in each country from the same organization. I I don't see this solving the economies of scale problem.

Could that affect proof of stake as well for a sharded stake thing as well? We could have val staked validators in regional subgroups.

Well, I mean the consensus comes from all of the stakers as a whole. So I don't think it really solves that either that way. Well, this is this is actually one of the other big reasons that you need to look at proof of work versus proof of stake is I have never seen a mechanism where you can reliably get sharded consensus in a stakebased system and however much you want to increase the size or the decrease the number

we have stakebased shed systems today right

at

multiple

how many scales of shards

I mean

and what is very large amount of scale

no how many shards And what is the latency between shards? See what ends up happening.

Yeah, the latency between shards is indeed the problem. Yes.

Right. Well, and it exponentially increases with the number of shards.

It's why I've moved away from sharding.

Well, but the problem is if you're in a single system, you're under the constraints of the internet and the physics of the propagation of light.

Well, I think maybe there I'm more of an optimist actually. I I actually have maybe a bit more optimism in hardware and software advancing. And the reality is we have single sharded blockchains that are keeping up with demand today and doing massive usage and bring in billions of dollars. There's something to say about that being good enough.

Is 2,000 TPS good enough?

We're doing a lot more than 2,000 TPS today in terms of capability if we measure it as like a onetoone value transfer. So like something like Salana for example's capable of maybe 40 to 50,000 TPS today. But that seems to be enough to fulfill demand. I mean, my theory on this, that's onetoone value transfers, right? Not complex transactions.

Um, but I mean, my theory on this, as long as blockchains can scale uh to keep up with demand, right? As long as we're not at demand, sacrificing speed for a sharded system just doesn't make sense from a users perspective, from a competitive perspective.

You misunderstand here.

Oh, right. Because because you don't make that trade-off in your system.

Correct. Effectively, so there's there's

Well, that's pretty cool. There's this argument of scale and and whatever argument you can make relative to what you can do in a single threaded system we can also do and and I would argue that you know proof of work is actually faster than proof of stake you know properly

what latency you guys achieving there

um in terms of

single block confirmation time

um so we get five to 8 second confirmations um

that's slow

but you can do pre-MPS in like 200

but we're not talking about pre-MP What are you talking about? This is fast. This is super slow.

That's a That's a dino chain. No, I'm sorry. Like 6 to 8 seconds is not fast.

What is What is the average time it takes for you to get a confirmation when you tap a Visa Mastercard?

Um, we're not you're not competing with Visa Mastercard anymore. You're competing with other cryptocurrencies that can do 400 milliseconds today.

I don't think you are. Ultimately, what you're trying to build is money. And I don't really see the other use cases as

So, you don't think speed is important? No, I think speed is important at the speed that money transacts and going past that is going past the case of money.

So in my view, in my view, good decentralized money needs good decentralized finance to support it. And if you want good decentralized finance to support good decentralized money, you need speed because you cannot build a competitive DEX if you don't have sub-second block times today. It

it's actually the opposite in my opinion. So

you're just acknowledge you're going very much against the grain now, right?

Oh, I go by first principles and that also oftentimes throws me against the grain.

I respect it. I'm there often too. Not on this one it seems. Now I'm the majority opinion it seems for a change.

So so effectively right what Salana does uh specifically is and Ethereum does this too is instead of mitigating mev you make it open so to speak. So you have it be part of the process of the chain and people who are transacting end up getting exploited.

We're getting a bit distracted here, right? Because now it feels like we're debating pure paralization versus sharding versus

No, no, no, no. Like like we're not

maybe maybe tie it back.

Yeah. So So you like you're saying speed is important and I'm trying to address this specifically in instead of just being okay with me. the the thing that you actually have to do to get rid of me is you have to slow your execution down below your consensus level on an economic bas.

Yeah, I don't think that's a good solution to sacrifice speed in order to eliminate me. I I would disagree with that trade-off, I suppose, but it's peculiar to me because you claim that work is faster, but Quay is not faster. Um it is it is more capable of doing higher transactions per second in a uh decentralized origination of those transactions with a similar number of nodes with a similar amount.

I would argue that's more due to sharding than it is proof for work.

No, that's specifically single threaded.

I mean we can do a million TPS with proof of stake sharded networks. Now why do we need this? um we we need proof of work because the long-term equilibrium incentives look like that of gold rather than

But this seems very handwavy. Like when you give us a philosophical sentence like that, it's not rooted in practical things like the actual measurable decentralization or the actual cost to attack. It's a philosophical nicity. It's a it's a it's a nice idea, but that like like we don't build an industry, we don't we don't take over the world with just nice ideas alone. Right.

Right. But Okay. So would would you agree that you don't need 400 milliseconds to buy goods and services cash and credit like properties?

Um yes, but my argument is that you need uh all of these other things like DeFi to be a success as a cryptocurrency. And the more fees you gain, the more secure, the more economically sound, the more scarce, etc. There's this flywheel in cryptoeconomics, right? becomes a virtual cycle and that's why I consider winning at at dex trading and winning at per exchanges is how you win as a cryptocurrency and if you neglect those things

then I don't think you're winning you're losing in my view

so so even if we disagree that that's a necessity it is still a an extra thing that you view can be done and more things is superior all other things equal right

yeah I mean I'm I'm very much an everything chain thesis guy I'm like everything has to go on the blockchain on the one good blockchain and that's how we get the most out of it right from a network effect and from an economics and a game theory perspective. So from the perspective of settlement, if we look at MEV, MEV in this idea is something that currently exists in finance, we call it dark pools. And anytime you do a trade, a market maker is taking money for you. And depending on your brokers and you have a zero broker fee, they're actually making all of their money by giving you poor execution through these dark pools, which actually pay for your order flow to frontr run you. Okay?

Okay, so if we're actually talking about what I think a winning DEX is long term, it's actually getting fair order execution when you trade assets rather than getting instant order execution that allows people to exploit your order. And so my point of view is if you look at what is practically used as money and what would make a good DEX long-term with fair execution, you actually have to slow the execution realization down to 15 20 seconds. But if you do do that, which is what you guys already experienced in your brokerage accounts, if you go to sell a stock and you place an order, you probably don't even hear back for 20 seconds that your order was.

We're living in a world today, right, where uh per exchanges, spot exchanges are making billions of dollars a year of revenue and have unprecedented amount of volume. Like to me, it seems like clearly this is what the market wants. the market wants fast transactions, fast trading, fast cheap swapping, etc. Right? Like that's clearly what's happening here.

And I can totally understand why that is the case from a practical perspective, why they would prioritize speed and and ease of use over what you're describing. It strikes me as so abstract. Well, I mean, you're injecting ease of use in there, but you're effectively arguing that they're wanting to give up one, two points so that they can have instant transactions.

But isn't that what the market but isn't that what the market is telling us? At what point would you consider yourself proven wrong? At what points in this timeline? Because I mean, you you're not going to keep saying for the next 10 years that slow Dex is the way when only fast Dexes are winning in this market, right?

I think we're starting to hit the end of the time. I've gotten a few glares from over there. Uh, so one sentence recap argument each and then we're out of here. What do you think? All right.

Um, so in my view, look, I love the cryptocurrency movement and I love decentralization. I love permissionlessness and I love censorship resistance. And I think the best way we can achieve that is by adopting the system that gives us the most security. It's by adopting the system that that gives us the most decentralization. You know, like you know, having a couple of large industrialists run like all of the uh consensus versus having a system where the stakeholders can run the consensus.

It seems to me very clear like which system is more decentralized. today almost in every there's not a single proof of work chain that can match some of the decentralization metrics we see in the top proof of stake chains today like that's just an objective fact right [snorts]

um sorry I shouldn't ask a question this is my conclusion right so look I I think that if we want to have the most secure and the most decentralized system then we need to look at this objectively and empirically and From that perspective, you know, there is no doubt in my mind that proof of stake is the superior system. You know, all of this narrative and and all of these stories we can tell ourselves don't change the facts whether we like them or not.

Okay, that was like 12 sentences, I think. Let's go, Dr. K.

I'll I'll try to do it in half as many. Uh so if if we look at the block size wars, I think that that time period was peak decentralization for Bitcoin. And some people look at that negatively, but in some ways I look at it positively because what Satoshi did in his design is he created a set of economic incentives that pit the holders against the miners and then inject the devs somewhere in between. And what 2017 really said was that you had a sufficiently decentralized set of actors so that the system did not want to change right it created oification and oification in the long term is what you need to create a decentralized system because I don't want to be dependent on governance. I don't want the majority of the wolves voting to eat the lamb.

If you create the same set of economic incentives in the money system that we currently have today, we would be foolish to expect that we will end up with a different result. And and that is fundamentally why I think proof of work is superior.

Thanks for the fun debate, guys. [applause] Could we have the big brains on stage? That could have gone another hour, I think. Yeah, she can tell they could probably go on for like, you know, 24-hour podcast. So, if you I think actually that's a good thing.

You know, you guys should go on a podcast for like, you know, 3-hour podcast. I would certainly tune in. So, I'm expecting to see that soon. So, um yeah, Eleanor, you can get on that. Well, um with that basically said, I will say um we've made it to the end of day one of Common Sense.

Um yeah, before I go, I do want to say something. I would like to welcome Anna Lord to the stage. Uh because she will probably have some some news to tell.

Um hi everyone. Um thank you so much for day one. [applause] Always a pleasure to bring together the brightest minds working on privacy, security, and freedom. I'm always so honored to bring this amazing group of people together. Uh we are here to build with intent, remind ourselves that the cypher punk values matter, that we are still there's still a fight to be fought, that we cannot let the institutional boring side of crypto win.

And I just talked about it in the in that interview. Maybe I am going against the grain and maybe this is completely crazy, but I feel like this team of people really believes in this and if we're going to dedicate our lives to building the future of money and finance, this is where it's happening with these people, with this technology. The cypher punk dream is not over. It's being built and we have to keep building it. So, thank you so so much for day one of Common Sense and I'll see you all tomorrow again with the hackathon finishing at 9:00 a.

m. judging and then the top 10 hackathon projects. Thank you so much. Good evening.