The Blockchain Debate Podcast

Motion: POS will overtake POW (Kevin Sekniqi vs. Tarun Chitra)

March 25, 2020 Richard Yan, Kevin Sekniqi, Tarun Chitra Episode 6
The Blockchain Debate Podcast
Motion: POS will overtake POW (Kevin Sekniqi vs. Tarun Chitra)
Show Notes Transcript

Guests:

Kevin Sekniqi (@kevinsekniqi) - debating FOR the motion
Tarun Chitra (@tarunchitra) - debating AGAINST the motion

Host:

Richard Yan (@gentso09)

Today’s motion is “POS will overtake POW.” 

This is an important topic for a few reasons. First, from an investor’s perspective, this is another angle of evaluating BTC vs its challengers, which by and large are POS, or moving to POS. Second, this provides guidance to architects of new protocols in making design decisions.

If we were to measure the dominance of POW via market cap, of the top 20 coins on CMC, coins with a POW mechanism (with no serious plans to migrate to something else), take up 74% of the share of market cap.

The two guests we have for this episode are: a builder of a highly anticipated POS-based protocol; and a creator of an extremely interesting business that simulates protocol behavior, to help protocol designers make more sound decisions.

I want to note that, about 70 minutes into the debate, the discussion shifted to a general conversation about money characteristics of bitcoin, and this goes on for about 25 minutes. This was not directly relevant to the motion, but I kept it in nonetheless, for the insights it shed.

Make sure to check out our previous episodes too, on Bitcoin’s store of value status, tokenization and smart contracts, DeFi, bitcoin halvening, and China’s future in blockchain.

If you would like to debate or want to nominate someone, please DM me at @blockdebate.

Please note that nothing in our podcast should be construed as financial advice.

Source of select items discussed in the debate:

Richard:

Welcome to another episode of the Blockchain Debate Podcast, where consensus is optional, but proof of thought is required. I'm your host, Richard Yan. Today's motion is: POS will overtake POW. This is an important topic for a few reasons. First, from an investor's perspective, this is another angle of evaluating BTC versus its challengers, which by and large are POS, or moving to POS. Second, this provides guidance to architects of new protocols in making design decisions. If we were to measure the dominance of POW via market cap of the top 20 coins on CMC, coins with a POW mechanism with no serious plans to migrate to something else, take up 74% of the share of market cap. The two guests we have for this episode are: a builder of a highly anticipated POS based protocol, and a creator of an extremely interesting business that simulates protocol behavior to help protocol designers make more sound decisions. I want to note that, about 70 minutes into the debate, the discussion shifted to a general conversation about money characteristics of Bitcoin, and this goes on for about 25 minutes. This was not directly relevant to the motion, but I kept it nonetheless for the insights it shed. Make sure to check out our previous episodes too on Bitcoin store of value, status tokenization and smart contracts, DeFi, Bitcoin halvening, and China's future in blockchain. If you'd like to debate or want to nominate someone, please DM me@blockdebate. Please note that nothing in our podcast should be construed as financial advice. I hope you'll enjoy listening to this debate as much as I enjoyed hosting it. This episode has a very high density of information, at least for me. Let's dive right in! Welcome to the debate. Consensus optional, proof of thought required. I'm your host, Richard Yan. Today's motion POS will overtake POW, or Proof of Stake will overtake Proof of Work. This is an age old question and I'm excited to hear inputs from our debaters. So to my metaphorical left is Kevin Sekniqi, arguing for the motion. He believes that POS will overtake POW. To my right is Tarun Chitra, arguing against the motion. He believes that POS will not overtake POW. Gentlemen. I'm very excited to have you join the show. Welcome.

Kevin:

Thank you for having us.

Tarun:

Yeah, thanks for having us.

Richard:

Great. Here's the bio for the two debaters. Kevin Seqniki is co- founder and chief protocol architect at AVA Labs, where they're building a new high-performance, layer-1 protocol with a focus on disrupting finance. He was previously a researcher and software engineer at various places including Microsoft, NASA, and American museum of natural history. Tarun Chitra is founder and CEO of Gauntlet, a simulation platform for crypto networks to help developers understand how decisions about security, governance and consensus mechanisms are likely to affect network activity and asset value. He previously worked in high frequency trading for Vatic labs and was a scientific programmer D.E. Shaw Research. As usual, the debate has three parts: an opening statements from both sides. Starting with Kevin. The second round is the body of the debate, with me directing questions to the debaters. Both sides are highly encouraged to follow up with their opponent after hearing answers on the other side. And of course they're also free to respond to each other's points raised during the opening statement. The last round is audience questions selected from Twitter and we will end with concluding remarks from both debaters. Currently our Twitter poll shows 59% saying POS will overtake POW, and 29% saying POS will not overtake POW. If the Twitter poll is representative of general opinion, then it sounds like Tarun, you're arguing in the unpopular and contrarian position. We will have a post-debate poll and whoever tips the ratio more to their side wins the debate. Okay. Let's get started with the opening statement. Kevin, please go ahead.

Kevin:

In regards to POS versus Proof of Work, obviously there is benefits to Proof of Work. it has demonstrated a new technology that you know, we didn't really know how to make use of for a very long time. Prior to it, there was work in distributed systems for decades, and we didn't really know how to deploy them in a really easy open way in pe rmissionless s ettings and prove work. Made it very simple. But Proof of Work to me is a stepping stone, a really important stepping stone in sort of getting the notion of permissionless network out there into the wild and really making us think about deploying financial value through pe rmissionless n etworks. but ultimately as far as a mechanism for sustainable longterm security and growth of these types of networks, it does not seem like a s u stainable mechanism. It is obviously incredibly expensive. As built, it leaks value to a very small set of participants with a lot of compounding effects and growth and economies of scale, that really make them outperform everybody else. So there's centralizing systems and, all in all, the set of arguments that have been used over time to claim that Proof of Work is the only mechanism by which true value is, you know, created base off a cryptocurrency or a sovereign currency, rather I should say, is a flawed argument, based on any measure of what market theory really means. Value comes from a general consensus from people that are willing to participate in a particular market, not from the underlying energy consumption of that technology or that system underlying instrument, rather. So that's my starting position.

Richard:

Okay, great. Tarun, time for your opening statement. Go ahead.

Tarun:

So while I agree that Proof of Stake has a really valuable place in the ecosystem and in the world, Proof of Stake and Proof of Work are dramatically divergent Sybil resistance mechanisms. And the main difference between Proof of Work and Proof of Stake is their functionality as Sybil resistance mechanisms, which ensure that the same person can't sort of claim increasing rewards relative to the resources they contribute to the system. So while on the surface we're usually able to liquidly convert capital to energy and backwards, it does seem like the security of Proof of Work and Proof of Stake are actually the same. However, during very adverse times, which are the times in which one really has to stress-test a Sybil resistance mechanism, it becomes much harder to view these two as the same. The natural diffusion of energy resources on the planet versus the generically concentrated distribution of POS assets makes it hard to imagine how, how sort of a POS system can, can really recover from say, a really catastrophic event. Like 99% of token holders lose their keys. In other words, there's sort of a weird thing of basically how you have to restart these networks. Joining them kind of effortlessly and permission-lessly takes a lot more work and you have to have sort of a significantly more complicated cryptographic setup. But outside of this sort of really doomsday scenario, the more pedestrian view is that financial attacks against Proof of Stake are far more effective than one might think naively. And one of the reasons for this, is that most Proof of Stake algorithms are usually analyzed under sort of a cryptographic threat model that is similar to the cryptographics that[inaudible] use for Proof of Work and other distributed systems. But this threat model sort of makes one secret assumption, which is that the capital in the system stays constant, and if there's a lot of volatility in the amount of capital flowing through these systems, you actually have very different behavior. Proof of work assets don't exactly have the same problem because they kind of need to also have hash power derivatives in order to have this sort of similar property, where capital or hash power, where security can kind of flow out of the system instantaneously. Given that there hasn't been much liquidity in these assets, it's hard to believe that Proof of Work security is really similar to Proof of Stake. The other thing is that the irreversible nature of burning energy versus the sort of inability to generate purely irreversible digital objects makes it hard to believe that Proof of Stake assets will really become dominant bedrock stores of value. However, I do really think that Proof of Stake has a really good place as a financial asset for something being more akin to US equities. Secure and definitely secure, but not really secure in the face of capital flight. And in the current times where we're watching capital flight happen, with, kind of, craziness that's going on in the repo market and different things that are happening in terms of flight to treasuries, we kind of are seeing the situations in which Proof of Stake security model is significantly worse than Proof of Work. Now at the same time, since we haven't defined what it means for Proof of Stake to win, or be better I will throw out one thing which, you know, a natural argument against, sort of, why Proof of Stake will not be subordinate-- Which is that US equities are significantly larger in aggregate and market cap than hard assets. So gold you know, other metals, oil. So while this is true the current financial system does still rely on the US dollar as the final settlement numeraire. On the other hand, the crypto markets, if you study market microstructure quite carefully, really relies on Bitcoin and Tether, and everything else is at this point a rounding error. So everyone eventually trades through those pairs, and the flight to safety that happens when a Proof of Stake token is attacked right now really still takes place via Bitcoin and Tether. And it's really hard to imagine that if there were sort of this, you know, a cryptocurrency took over the world and was mainly used all over the world, and if there were a sort of scare, kind of like the coronavirus scare, it would be kind of hard to imagine people actually sticking in a Proof of Stake coin, when asset prices are going down and capital is being reallocated. So that's sort of my view on this. It was more that Proof of Stake presents a bunch of risks that make it not suitable as hard money, but they make it much more suitable for financial instruments for lending and equity, which is still extremely valuable of course.

Richard:

Okay, great. So let's proceed to round two, which is the body of the debate here. I'll be directing questions to each of you, and feel free to respond to each other as opening statements. My first question is for Kevin: Paul Sztorc, from the die hard POW camp says this:"Nothing is cheaper than POW because, as an economic principle, cost of production will always catch up to revenues. So even if POS doesn't directly waste energy, the competition amongst validators will still drive up the costs eventually approaching the block rewards." His argument is that if the market is efficient, then the cost of mining, the cost of validating will still match the revenue obtained. In this case, you might not be burning energy directly, but you'd still be consuming a high amount of cost. So do you have any thoughts on this?

Kevin:

Certainly both Proof of Stake and Proof of Work are Sybil deterrence mechanisms. And naturally they need to be costly. And that's their whole point. Sybil control mechanisms need to prevent a set of participants to come into permissionless system and overtake it, and double spend and do all sorts of malicious things. The cost of attacking a Proof of Stake system should be very high, especially as the system grows with its active usage. That's just a natural consequence. I mean, that's exactly what you would want to happen. So yes, I mean I do agree that the true costs of running a highly secure Proof of Stake system is very high, does stand, and you know, it should be as high as the cost generally in US dollars as running a Proof of Work system. But the arguments on costs are hiding a very important factor, which is on how these costs are derived and where these costs are being distributed. The security model here is significantly different, even though it aims to achieve the same thing. Proof of work leaks to a small set of participants that you know, uses economies of scale to grow larger and larger using their hardware equipment. The cost there is in this massive investment in hardware, whereas the costs in the Proof of Stake is in capital. So yes, both very costly, but it seems like a just not fully formed, at least, argument to say that they are both as inefficient. In fact quite the opposite. Proof of Stake is probably one of the most efficient mechanisms of ensuring security in the system, because ultimately both Proof of Work and Proof of Stake try to attach a cost value to it; it's just that Proof of Work does it through a very roundabout way of, you know, first bringing in hardware and so on. Whereas a Proof of Stake directly ties it to capital. So it's much more efficient. It removes any extraneous exogenous mechanisms of bringing in security such as electricity costs and mining hardware. So it would really need to be a much more formed argument on his side in order for me to really attach to it. but that's sort of the very gist of the answer.

Tarun:

Yeah. So I directionally actually agree with Kevin, that it's actually really hard to correctly measure this sort of delta in production costs versus revenue costs, because they kind of break down into quite different components that are not quite convertible. And in a lot of ways you know, Sztorc's argument does take sort of a more heuristic approach to justifying this. There are more formal ways to do this, which is that you can kind of try to consider a portfolio that consists of a Proof of Work coin and a Proof of Stake coin. And then you can also consider a portfolio of other derivative securities that are trying to replicate them, as in having the same exact return profile and the same volatility profile. And you can basically roughly speaking, show that, if there are hash power derivatives, the correct way to sort of replicate a Proof of Work portfolio with a Proof of Stake coin is to basically go long a staking coin and short a staking derivative, which will be equivalent to a Proof of Work coin plus a hash power derivative. And from this type of replication argument, you can kind of use traditional portfolio theory to kind of estimate this Coase-an tax, which is the delta between costs of production and costs of revenue. And it's kind of really hard to argue directionally, in my mind, which one is actually bigger, although I do inherently think it's probably Proof of Work. However, the costs for Proof of Stake I think are significantly higher than people had assumed, I think during the boom times in 2017. part of the reason is like, you end up running a lot more nodes than you expected. You end up being in a data center that usually has to have a very high bandwidth interconnect. You need to have DDoS prevention and[inaudible]. You need to have key management for all of these new curves, which means you need a lot more complicated HSMs. You need a lot more new technology that hasn't been proven in production yet. And these temporal effects will definitely cause some type of Coase-an slippage, and Proof of Stake in which you will still have these concentration effects into certain professional groups. And we already see that. I just don't think that the Coase-an tax is quite as large as the Proof of Work one. Because the Proof of Work one does have this weird thing where the timescale in which you have to lock up capital for hardware is probably a lot longer, unless HSM prices start going through the roof.

Richard:

Sorry. Are you able to clarify that concept that you mentioned, Coase-an tax?

Tarun:

Yes, sure. So, Coase-an tax traditionally in sort of the... Coase has this very famous paper on the Theory of the Firm. The idea is that why do firms exist? Why, why can't everyone just be a contractor? What are the sort of benefits of forming groups and firms? And part of the reason is, if we think of the network of agents participating in an economic network, if they're interacting at a peer to peer level only, they end up maximizing their own local revenue. Then they might charge someone who they need to work together with collaboratively the maximum price they can charge them. Whereas when you have a firm, and you have a group of people who interact while charging each other no transaction taxes, the firm as a whole can grow faster than the sum of the individuals. And it's sort of a network effect-like thing. And so the Coase-an tax is sort of the measure of the difference of"I take N people and I put them into a firm, and then measure the total output that they have" versus"taking N people as contractors, more or less, and measure their total value accrued." And that sort of difference is roughly speaking this Coase-an behavior. Now, the reason I mentioned that for the cost of production versus cost of revenue is there's a sense, in which the difference between those goes to zero when there's a lot of competition. And there's no, kind of, impediment to starting a new firm that's maximizing value. So roughly speaking, you could kind of view the Sztorc argument from that lens if you're, if you're taking kind of traditional economic theory.

Richard:

Okay. Kevin, you can respond to this, or the other points that Tarun mentioned in the opening statement.

Kevin:

Sure. No, I absolutely agree with Tarun here. the thing is that you know, something that we need to do as the debate progresses, and I think I'm noticing this, is that there is a multidimensional debate happening here with Proof of Work and Proof of Stake. We're treating it as almost, you know, apples to oranges comparison. But it's actually a very complicated set of technologies that interplay differently based on a specific implementation of each one of these. So, as we go forward, I would want to tackle on a particular functionality or a particular property of Proof of Stake or Proof of Work, as it pertains to a particular implementation of each one of these because it's really complicated. It's very multidimensional here. In regards to the points that Tarun made in his opening remarks. So when we describe the two, let me take the approach of first, sort of, talking about them in the sense of just, raw sustainability-- the raw number, rather the overhead of running the systems. And from that perspective, you know, how much electricity is being burned, how much are we actually expanding, how much are we using to secure the system? It's kind of very hard to make the case that the Proof of Work is in any way sustainable, or will be the technology that will win out in the long-term. But I do see the point and in fact, you know, we agreed off band at some point on our discussion that indeed, you know, Proof of Stake as has been deployed currently in the existing systems that at least I have seen, has this weird property of a quick capital flight. So you know, there could be some black Swan event which could disrupt the value of the system, causing a rapid move out of it and therefore destabilizing the entire security guarantees, and any applications built on top. Not quite the same can be said for Proof of Work. However, it kind of can, because Proof of Stake, something that people fail to realize, is that it can almost fully emulate Proof of Work in every regard as far as properties go. and vice versa. In a lot of ways. It's really just about the exogenous costs that are attributable to each one of these two mechanisms. Ultimately, they can replicate each o ther's properties very closely. So while there can be capital flight in a Proof of Stake system, people can-- you know, for example, let's s uppose t hat there is a DeFi application and the DeFi application all of a sudden is a very attractive investment for many s takeholders in the system, causing validators to move from actually s ecuring t he system to just locking their coins into the DeFi application. This could cause a very rapid flight of security. The same can be said for Proof of Work. Of course, you know, it would be the case that everybody in a Proof of Work system is all of a sudden turning off their machines. The hash power drops radically but it's effectively the same property. You have a massive capital flight even though in this case it's hardware flight, not really c apital, and that hardware flight does in fact reduce the security system pretty drastically. The point that I will concede to Tarun's advantage is that it's a more coordinated type of event in Proof of Work, or rather, maybe more of an even rare black Swan, in that you would need for let's say, Bitcoin to have global consensus on pulling the plug-- you know, worldwide, e verybody i s turning off their machines in order for such an event to happen, whereas we c ould never really predict you know, and we have kind of actually seen this happening. In Proof of Stake, this can happen pretty easily. We can see it happening. It's much easier to believe that a social event can happen because, you know, we just suddenly moved to a really highly attractive DeFi application. But ultimately, you know, as I said in the beginning, the two systems can actually emulate each other n early perfectly. I don't really have a universal composability theory of Proof of Work a nd Proof of Stake showing that both of them can emulate each other. But you know, you can make capital flight harder. I would say on Proof of Stake by you know, forcing that Proof of Stake or ra ther s taked coins in a Proof of Stake system are locked for very long periods of time, in the same way that it would require somebody to buy a lot of hardware coming into the system. And then if they'd want to leave, sell all that hardware back on the open market, that same sort of overhead can be emulated pretty closely in a Proof of Stake system. So from a security perspective, and I'm not even talking about the other arguments that people have erroneously attributed to Proof of Work in regards to its value attribution, in regards to just the security, that particular concern does not necessarily seem like something that cannot be fixed. Although to concede it has been overlooked wildly. And I think Tarun has done a fantastic job at precisely identifying these risks with current Proof of Stake systems, and how maybe easy it is to observe capital flight.

Richard:

Okay. So Tarun, would you like to respond because I actually have a few points myself too, but you should go first if you have immediate thoughts.

Tarun:

For sure, yeah, thanks. I think the main difference is really the timescale difference. So if we think back to the history of Proof of Stake, I guess the first real mention of it was that, that I could find, I'm sure there's probably some other place where this happened earlier, was a 2012 BitcoinTalk post, where, you know, someone did the thought experiment of-- what if I was doing Proof of Work mining, but I could every for every block reward I got, instead of having to like sell the block reward, buy hardware, put the hardware online, increase my relative hash power, what if I could just reinvest it instantly? Like virtually like the network was able to take my block reward, and I say, hey, this block reward I just earned, actually just make that increase the probability of me winning the next reward. That experiment of like doing this continuous reinvestment, this instantaneous reinvestment, changes the timescales in Proof of Stake, such that the risks are always dependent on the fact that you have this like constant reinvesting property and you can prevent it by adding lockups and adding, you know, other forms of mechanism for doing things like that. But that's really, that makes the network very hard to bootstrap. And it's actually oftentimes quite capital inefficient for participants, especially if they need to, you know, pay costs in dollars for validation hardware and you know, DDoS prevention, et cetera. And HSMs and updating HSMs. HSM meaning hardware security module. So that's the thing that you can use to store your keys that usually will evaluate, say certain elliptic curves. It will, in the case of like ZK-proof, it will evaluate a pairing, et cetera. They, you know, provided that those costs are still denominated in another currency. That's your numeraire. It's very, very hard to imagine people viewing kind of the capital efficiency of Proof of Sake as being optimal now. Now can it be fixed? I don't doubt that it can get better. I just am not, I just would never, I wouldn't be willing to wager a really much more than a 50/50 bet on whether it can come within, you know, 20% Proof of Work as a capital efficiency, from a capital efficiency standpoint.

Richard:

Sorry. But isn't that capital inefficiency sort of mirroring between POW and POS, in the sense that if you were to purchase a mining machine, you're also sort of tying in your capital and then there's severe friction in you selling that machine to reclaim your capital. So doesn't that sort of mirror the long capital lockup characteristic that Kevin posited in POS?

Tarun:

Yeah, there is definitely a sense in which that's true, although there's a more clear methodology for amortizing the cost of your opex and capex in mining, versus the fact that in Proof of Stake right now it seems like opex is not zero for sure. It's definitely very far from zero. Otherwise people could validate on their cell phone. But it's not a, you know, greater than 50%. That trade-off for opex and capex at least in Proof of Work because of how liquid the energy derivatives market is- you know, it's one of the most liquid markets in the world outside of the rates markets- the energy derivatives markets help you hedge and amortize your opex significantly better than you can in Proof of Stake right now because there's just not that whole ecosystem of hedging products. And so that's why I'm saying, you know, if I were to start being a validator today, that my capital efficiency is actually quite limited by the set of financial products that I can use to kind of achieve sort of what I want to be an optimum outcome. Whereas in Proof of Work, that's not true because e verything i s structured around these energy financial products, which are extremely, extremely liquid compared to anything in crypto.

Richard:

Okay. Okay. Alright. So let's move on to the next question. It's for Tarun. So one way for Bitcoiners to combat the whole energy wasting accusation is pointing out the fact that much of the energy comes from energy sources. There are excessive or hard to store. Examples are solar, wind, hydro. This type of green energy for mining analogy for staking is that if an ETH HODLer wants to hold ETH regardless, then they're not taking up additional capital resources. Do you think this view is legitimate?

Tarun:

Yeah, I think there's one kind of potential difference in HODLing in Proof of Stake networks, which is, and is the moment that you have raw staking derivatives. The HODL'd ETH is actually in the insurance fund for the staking derivatives. So what does this mean? So if we unpack this a little bit, the staking derivative is sort of letting you borrow kind of like a Maker CDP against your staked tokens, some fraction: let's say it's 30%. So you have a hundred staked tokens, you can borrow 30 of the staked tokens and that mints a synthetic that's supposed to represent that. If you get slashed for 30 staked tokens, then your collaterals are reclaimed by the network. But the network is socializing the loss and the ETH HODL'er in that case is actually potentially, is sort of implicitly providing the insurance fund for that. Now, I'm not saying that's necessarily good or bad, I'm not taking a strong stance on whether it's good or bad to to socialize those losses to holders and in those ways. but I am saying that it is, it is something that you kind of have to consider when you, when you say, Oh there is a you're not really taking up capital resources, you are providing this kind of re-insurance to the rest of the network and you might not be correctly assessing that you're doing that in a risk neutral world. You are taking no capital resources but in a somewhat risk averse world you actually are taking up quite, quite a bit of capital resources under certain types of black Swan events. and the other thing that's kind of weird about the ETH HODL'er HODL'ing ETH regardless is at least you know for a lot of the Proof of Stake networks that are going to launch, they have emission schedules. So block reward and transaction fee redistribution schedules that are dependent on the total stakes quantity. And so by being an ETH HODL'er who is not staked, you know, you are getting burned by inflation, but you're actually getting burned by inflation significantly less if everyone else is staking. So if a lot of other people are staking, let's say, let's pretend the curve is at 10%, the block reward is 10 ETH, 10% of outstanding ETH staked. The block reward is 10%. And when it's 90% staked, the block reward is 1%. Now there are very few people who are traders who would not take a liquidity premium, especially in cryptocurrency of 1%. Right? And so you're implicitly, actually as an ETH HODL'er doing better than the stakers at really high stake rates. Now you might argue that's good or you might argue that's bad, but these are the types of risks that you're implicitly taking.

Richard:

Okay, great. Well, Kevin, do you want to respond to that? If not, I'm happy to move on to the next question.

Kevin:

No, no, certainly, absolutely. I think this is a straightforward one. Totally agreed on the characterization of the current effectively Etherium ecosystem right now. But as I said at some point that this i s a multi-dimensional, you know, argument and it needs to be framed in the context of the particular system that is implementing a Proof of Stake Sybil control mechanism. So what Tarun is mentioning as e ffective w ith these being put options for future c ash flows, absolutely right. But this can be easily remedied. Obviously you can imagine, for example, a Proof of Stake system where there is a sort of two-coin dynamics. There is the same coin which has, you k now, a behavior that makes it look more like hardware, which has an onboarding period, a d eloading period, being sort of quote unquote, sold on the open markets, that basically precisely mirrors what, you know, regular mining h ardware would look like, and the ability for these tokens to move back and forth from one to another based on some bonding and un-bonding period. So I see this, no, I don't have this worked out mathematically fully. And this would be a really interesting way to go forward. But this is simply a failure of current implementations, not necessarily a fundamental failure but as Tarun has done a fantastic job identifying this current failure and in fact pretty surprised that people weren't seeing this, sort of, from the very beginning. but it's a really great and id e ntification of the problem. but I see pretty stra ightforward rem edies to it, and make it really emulate the same exact way that that you would operate with your capi tal if you were to go down the mining route.

Richard:

Okay, great. My next question is for Kevin. Again, POS will result in lockup of huge amount of capital to secure the chain. This capital is inaccessible to the broader economy and therefore cannot be used to facilitate technological progress and economic growth. So this is another point raised by Paul Sztorc in one of his articles in favor of POW. So your thoughts on this.

Kevin:

Effectively what he's saying is that there is a portion of the capital broader pool that is being used to secure the network, and the rest is out there, floating in the network. Well, like these things are divisible, infinitesimally divisible. it sounded like we have these coins that, you know, there is a fixed supply of them or whatever. It may be and they're not divisible. Therefore, these things are never, you know, we're fighting over just a very small finite set of tokens and therefore removing some of them from circulation by sticking them caused us a huge panic on the market. I don't think we've seen this empirically, anywhere both on crypto and non-crypto markets. This is just not the case. In non-crypto markets, may be, there might've been some instances because these things are not infinitesimally divisible. But in crypto markets, e ffectively that's the case. hese systems, it's just very hard to, to make the case for what Paul is saying.

Richard:

Just to clarify on that point though, I think basically, if you were to compare two systems, a POW one and a POS one, suppose they have the same market cap and same level of security, equal unquote, I don't know exactly how to quantify that, but let's say in broad strokes they have similar amount of security. It just seems that the amount of capital that goes into funding the POW system, the opex is the ongoing electricity burning and the capex is the money equipment. And then on the other hand, the opex, I suppose is the opportunity costs of the capital lockup, but the capex is substantially all the capital that the POS holders are willing to put in for staking. Right? So does it make sense to say that the budget used to be securing the system from a capex perspective is much larger than the former? And if so, then I think that's where Paul's argument sort of kicks in and that just says you're spending way more capital to provide the same level of security. I'm not saying that's a valid point, but I'm just clarifying his stance and I was hoping to maybe hear your opinion on whether this stance is legitimate.

Kevin:

Absolutely. So right, exactly as you said. It's just the problem sort of basing you know, that argument on the first set of assumptions seems really difficult to make here. we haven't seen this empirically. it seems really hard to make the case. And conversely, if we're going down this route, Paul is failing to mention the fact that, you know, all the new capital that is being generated, or rather all the new coins that are being minted are effectively being paid immediately out to the same people that are se curing t he system, which can of course take them out of circulation, can do whatever they want with these tokens. So you know, a lot of the arguments here seems, by the way, also the argument that I'm currently making is not necessarily strong. It' s ju st an assumption. So it seems like both sides are really starting with assumptions, and not clear based in empirical evidence.

Richard:

Tarun- Would you like to follow up on this?

Tarun:

Yeah, so I think you know, the capex, opex version of the world is the correct way to frame this question. And one of the reasons, while I think Proof of Stake will work and will be useful, I'm not bullish on it having the same capital efficiency, strictly because the Proof of Work opex market is very closely tied, again, to the energy derivatives market. And in some sense the opex is a lot, can, like, track realistic demand and energy usage much more closely, I think, than, currently the opex in Proof of Stake can. And part of the reason is, you know, as you pointed out, one component of opex in Proof of Stake is opportunity costs-- lost yield from an external source or, or if you're, risk seeking maybe a higher risk thing; if you're risk averting, then a lower risk thing. But there's another component of opex that I think is more, was not really expected when people were first building these systems, which ended up being the kind of bandwidth requirements having to run nodes and in quite good data centers, right? Like a lot of validators, professional validators are running in, you know, Equinix databases, at Equinix data centers, which was like, that's where NYSE, and, and you know, the CBOE keep half of their data centers. And so it's like there's a sense in which there's a lot more security that you need at the network level right now, which again could ameliorate over time. Combined with this idea, and so that causes a bunch of opex at least right now that's not just related to the opportunity costs of capital. Combined with the fact that there's just not a really good way to hedge, other than sort of providing some kind of loans against your state to quantity, but that has a ton of security effects. And so that's why even though you've seen all these validators make a lot of proclamations about wanting to have staking derivatives and wanting to have them, you know, be handled by the protocol so that the consensus mechanism could like at least kind of stem capital flight. These are still very early ideas versus the opex management in the energy market being extremely, extremely liquid and optimized. I'm not saying that this can't change, I'm just saying that you have to reinvent a lot of existing finance, that I would say that people in the cryptocurrency space just don't totally have, which in some ways is good. You're starting from a new shoe, but in other ways it's bad. You're going to redo the same mistakes as, as Matt Levine always says. So I do agree that maybe in the sort of really long time limit past my lifetime, this may actually work out, but for the foreseeable future, it just seems really hard to imagine the capital efficiency on the PR for Proof of Stake improving, like having a Moore's law.

Richard:

Let's shift gears a little bit. My next question for Tarun is: One advantage of POS over POW is the punishment mechanism. In the former system, the offender's stake can be confiscated through slashing and so forth. Whereas in the latter system, the offender just wastes the power being generated for that one block. So it just seems that there is a lot more flexibility when it comes to the POS design in terms of keeping the stakeholders in line. Your thoughts on this?

Tarun:

Yeah, for sure. I think this is probably the place where Kevin and I will have some more disagreements on, which is good for a debate. But I think that there is a really huge advantage to having irreversible stochastic processes-- this pseudo Markov sampling method of Proof of Work driving your consensus mechanism versus potentially reversible methods. Because that's really what makes financial attacks a lot harder. And so part of the reason that I think Proof of Stake has such a way more interesting and rich design space is because it does kind of have these the financial surface area of like, you know, you're something that's supposed to be a store of value now has a financial surface area of, you know, a CDS on sovereign nation. And so that has a lot more complexity and like the end-user has to know how to deal with that. And so that, that also happens during slashing and, with confiscated sort of collateral. But there's also a particularly weird incentive, that is weird for Proof of Stake that doesn't exist for Proof of Work, which is that slashing can be sort of insured. So again, I get back to, to staking derivatives. If your network has staking derivatives and you think you're about to get slashed, you should be taking out a loan against your stake quantity, the largest loan you can and then just defaulting, getting slashed until it defaults. But you've already sold it for some stable coin and now the network has to eat your loss. And while you think this might not happen you know, Compound and Maker, which are our most, which is the closest we have to battle tested production systems of this form. basically already have run into this, where there are some of the less liquid markets have had people borrow stable coins against more or less coins, and they basically default on their loan and no liquidator wants to actually even buy the collateral, and kind of make the contract solvent. Staking derivatives interactions with slashing, will have a similar kind of like moral hazard and I think burning energy can't be accidentally burning energy, can't be insured and as cleanly as this. And so, you know, the best you can do there is to try to hedge, you know, the aggregate loss that you have to, to smooth your income out or to join a pool. Whereas in Proof of Stake land, it's a little more like, Oh, well I could theoretically insure my slash rate and then just default on the note. Now I think when, when you don't have slashing as Kevin, I'm sure we'll have a lot to say about, you do avoid this problem, but you do still have like complicated, long-term payoff. The amortization that like I just really hard to imagine replicating Proof of Work. Now this is why I say I think Proof of Stake is a great financial instrument platform. Is it money? Absolutely not. Is it a real store of value that should be great for collateral? No, not really. Like there's just the surface area makes it so much more prone to these kinds of, these kinds of like weird insurance style attacks. And the space of the, I would, I would posit that the space of derivative securities on Proof of Work assets is actually finite dimensional and quite small. Whereas the space of derivative security is on Proof of Stake assets is infinite dimensional, which is why they are good for programmability, but also why they're kind of second tier, they're the mezzanine debt to say, you know, the top tier debt. And you know, I wrote a little note when I saw, kind of heard this question that was you know, it's maybe a little bit of a wonky note, but you know, reversible computers which had been talked about since the 1960s, which are super cool, right? They use really crazy physics to generate kind of totally reversible computations and there's reasons that reversible computations are more efficient theoretically. But in practice, the computer we're using right now, the computer you're using right now uses irreversible thermodynamics and burns a lot of energy. And that's just c ause t hey're just easier to productize o n. The UX ends up being easier. and you don't have to think as much. Reversible computers have k ind o f the same surface area problem. so that's sort of, that's sort of my muddled set of thoughts.

Richard:

Fantastic. So Tarun's opening statement was quite a bit longer, and I just wanted to make sure that there were points in there that Kevin had a chance to respond to. Tarun, maybe, if you feel there are certain concepts or arguments that have not been fully hashed out through this discussion, feel free to bring them up. Maybe summarize them for our benefit because, to be honest, I think your arguments are quite dense.

Tarun:

Yeah. I think, again, a lot of, a lot of the points that Kevin has made have really covered things. I think the main thing that especially that we've seen in production lately is the increased usage of flash loans in the device space, and flash loans are really kind of this magnificent application of cryptocurrencies, and blockchains that just did not fundamentally exist in the normal markets and does not really have an analog. There really is no way of doing it without trusted, without trustless computation. And so, I mean, I think they're fantastic from this point of view, of they've expanded our notion of knowledge. They've also changed the necessary threat model for capital based systems like DeFi and Proof of Stake where the capital is a digital asset. And so I think one thing that, you know, especially as Proof of Stake grows that people will have to think about is that the threat model is not the 50% or 33% attack anymore. It's that someone can instantaneously get 99% of the network for a tiny amount of time. And so in Proof of Work, it's just impossible to imagine doing that unless you have hash power derivatives that are liquid. And we've just seen the market for those be quite poor. The empirical evidence is that people don't really want them. They would rather hedge directly in Bitcoin or hedge in Tether. so I think that, we're kind of in this amazing, amazing race of how can I destroy these networks as fast as possible in Proof of Stake land. because the surface area is gigantic. And I think trying to figure out like, you know, I would love to know where Kevin thinks, what the timescale is. He thinks that kind of we will have a manageable understanding of the surface area and sort of when it can approximate the kind of threat model of Proof of Work.

Richard:

Kevin feel free to respond.

Kevin:

You know, Tarun is obviously one of the people that I respect most in the space, but I think he's also over complicating things. It's actually pretty straight forward. The argument that he's making is that, effectively, Proof of Work has exogenous threat model to the capital that is being used and programmed inside the blockchain platform. Very simple. There is a set of monies, these are programmable. they can be used for various different functions. Proof of work says, I am going to bring in security without really touching the internal c omponent. Money itself. Proof of Stake says, look, I can use the money that is currently in this blockchain to itself encode security and make it hard to attack, et cetera, et cetera. So Proof of Stake is a much larger set of it's a larger family o ffer of protocols, which one of those could be e ffectively t he simulation of the function underlying the Proof of Work mechanism. You can in fact fully emulate Proof of Work using a Proof of Stake system. This is something that I think Tarun is missing entirely. Which is then the sequitur here is that yes, you in the current designs, and I think I've mentioned this before in very straightforward terms, the current designer Proof of Stake systems is that it intertwines the security of the system with the actual capital that is being used for various other function a little bit too much, but that can be easily mitigated or reduced down to basically the[inaudible] of error probability by increasing bonding times, and by making maybe even something like a two coin system where there is a coin that is being used. Or rather, part of the ecosystem coins are specifically being used in this new sort of ionic format, you know, that do not really interfere with the rest of the system, and can transition or metastasize into something different. Rather, the other capital that is being used in the system through some process as it takes some period of time that exactly emulates Proof of Work. So the entire set of arguments here on, you know, slashing risks all these things really b oiled down to the fact, to the f ollowing fact, really b oiled down very simply, current Proof of Stake designs intertwine the c apital that is being used for all sorts of non-security based systems, with the c apital that is being used for the s ecurity o f the system. In fact, it happens to be an in current Proof of Stake designs. These two forms of capital are the same, the one and the same. But these, there's no reason for this to be the case, right? It's just, it was the simple way to construct Proof of Stake systems. We went with it. And now we're discovering that as you know, I think Tarun is pioneering, this design is probably not the right approach. And we need a way to transition between these two very separate classes of coins, yet there is no reason why we cannot assign value and security with the, with the very coins that we're programming. So we can use those coins, that programmability that the coins provide you, and tie to the secure the system. Proof of Work is simply, you know, a first attempt at a, rather, mapping hardware extraneous hardware to the security system and security through that way. But all of the things with capital flight, de-stabilization, black swans that apply to Proof of Stake can easily apply to Proof of Work in the same exact way. So I don't particularly see a difference in the two. But I will concede that current Proof of Stake designs could be done much better.

Richard:

Okay, great. And by the way, speaking of slashing, my understanding is Ava does not have slashing, is that correct, Kevin? And what's the thinking behind that?

Kevin:

Oh, it's very simple. It's because we don't, we want to isolate the capital that is being used for staking from the rest of the network as much as possible. Staking seems to introduce a deeper sort of interlock between the two, because you're effectively saying that this is the same capital and you're going to slash this capital. So you better not do it. We're taking more of the Proof of Work type simulation approach where there isn't quite a two token system in Ava, but there is one type of token which is pretty aggressively different from the other set of tokens which are used for all kinds of you know, capital intensive applications like DeFi applications. So we can't really intertwine the two. And when you don't intertwine the two, then slashing seems just like another thing that you can get rid of. And that's great because it's something that increases complexity that you really don't want to go down that route.

Richard:

Okay. Got it. Got it. Actually, one more question for Tarun before we move on to audience questions. So Arthur from Tezos and Gün at Ava asserted that in POW transaction fees accrue to the hash power of the network with no corresponding reduction in inflation. On the other hand, in POS, fees accrue to stakers, which effectively lowers real inflation. So the idea here is that POS does a better job of value preservation, whereas the POW miners seem to have no skin in the game. So your thoughts on this.

Tarun:

I mean, maybe the more empirical view on this is that we just haven't had real transaction demand to like transaction demand where the term transaction demand was dominating the block reward subsidy to actually know whether this would be true or false. I think I definitely think the naive version of that is very true. The only thing I would say that I would add that might, I would disagree with on that is that the potential for validating pools to actually take the fees like to appropriate a higher percentage of the fees then of the block reward subsidy I think is quite high. In the same way that right now people are kind of suggesting that miners should be front running all the flash loan transactions, that are profitable enough if they could like figure out how to do it quick enough. I tend to think that the fact that we're coming to a world where the, at least it seems right now where there are mainly professional, some set of professional validation services who actually manage the Mempools and who actually are running the nodes. I just suspect that they will find a way to front run transactions that have high value when they exist, which may make their, kind of, like, skin in the game a little bit weird because they're sort of custodians who are front running their own delegates and users of the system. But this is a very, kind of like, nuanced thing that like I said, we haven't seen real transaction volume on any of these networks including... Like Bitcoin is all, you know, if you look at the transaction fees, the really are, this like, mean reverting process except around big price moves. Like certain types of big price moves out. Other than that, it looks almost like noise. There's very, there's very little information content in the fees for I think making an empirical version of this statement come true.

Richard:

Okay, great. So let's move on to the last round of questions from audience members from Twitter. So these questions are for both, and feel free to answer in whichever order you prefer. So the first question, this is more of a comment, but I'd love to get your reaction to it. So this is from somebody named last token. POS is great for consortium or oligarchy-ish situation. POW is fit for extreme decentralization, and egalitarian situation. Agree, disagree, merits?

Tarun:

I mean I definitely agree with that sentiment. That's why I said I think Proof of Stake is really great at being US equities, which isn't oligarchy. Let's establish that that is a consortium. It's a somewhat big consortium. There are a bunch of exchanges. Reg NMS from the SEC kind of forces them to work cooperatively to some extent. They don't destroy, try to completely destroy each other. But I still just don't know if I can, I'm still just not convinced that it is equivalent to money. Or like, like under this extreme decentralization of I want to convince someone who like, doesn't know, understand what an option price is that like they should accept this thing as a payment. Now I'm not saying any of the current POW coins are; obviously that are not that either. I just, it's just harder for me to imagine going to my grand-mom and being like, okay, so here's the vol surface you got, the thing you're buying. And you're buying at this point. And like, it might go to this point, but it probably won't. Like, it just seems, you know, maybe there will be a way at some point where we really know the security of these things, understand the security in such a way that the UX is, feels the same. But I do think that like, you will end up having sophisticated groups of participants being very good at Proof of Stake, and Proof of Work is kinda good for unsophisticated participants. You don't need to be that intelligent to figure out how to do it correctly. Right now. At least that's what I'd say.

Richard:

Kevin, anything to add?

Kevin:

Yeah, no, definitely disagree on those fronts. In fact, quite the opposite on, on every front. Proof of Work is probably a system which because it has separated the capital from the security, the people that are providing the security don't necessarily care about the capital that's inside the system. They simply get these rewards, they can dump on the open market and they can pull the plug at any time. That's A. Two, it is a highly centralizing system. It incentives for cooperation and centralization into large pools. And objectively speaking, the empirical results demonstrate exactly this. And C, I think I've mentioned many times, is that the Proof of Stake system can just emulate, or rather can because it's so programmable because the security itself comes from the, from a programmable process. Then and I mean like, yeah, code wise can be encoded in any way you want. Then you can really emulate all the processes of Proof of Work directly i n a Proof of Stake system. Of course, I will concede that Proof of Stake systems that have been deployed to date are pretty centralized. They are basically running on a few validator services. But you know, that could be mostly because of the s o rt of immaturity of the Proof of Stake deployments. They are really being deployed on just a few highly trusted effectively validator-as-a-service providers. And that kin d of ma kes sense as the ecosystem evolves, evolves further. I wou ld su spect that the offerings would increase. but we'll see. It's very possible that we' re ma ybe centralized ev en further on the Proof of Stake offerings. So that's definitely a concern. We would want to see people deploy widely across jurisdictions and really across the world. But you know, the argument here that this isn't money, it's just simply, you know, there's, there's no argument for it. Process-wise, taking a look at this as this functional process, both on the Proof of Work side, on the Proof of Stake side, it's very clear to me that the latter can fully replicate the former so whatever the latter is, can be made possible in a properly designed Proof of Stake system.

Tarun:

So I actually have one, maybe two comments to that. So, so one is you know, if you ask people to put their money where their mouth is, my, I believe in the, that Proof of Stake systems, some of them will, will definitely survive, but I own zero Proof of Stake coins. The only thing I own is Equinix stock. And I own a lot of Equinix stock, mainly because I think the data center providers are going to benefit quite, quite a bit, more than the, in some ways potentially more than the coin themselves as a from a risk adjusted standpoint. The second thing is I do agree that technically Proof of Stake networks can emulate Proof of Work, but my question is really more around the complexity of it. Like, you know, some of 5 million mile version of it, Can Proof of Stake emulate Proof of Work in polynomial time. And so that's actually that question itself. Given all of these, as we keep finding all of these new risk adjustments and new things that we need to adjust for, it's not actually totally clear that the computational complexity of Proof of Stake emulating Proof of Work is actually sufficiently small, that we can emulate it with a perfect approximation ratio of capital efficiency of 1 relative to capital efficiency of the other. If that's true, if that turns out to be true, and it's not even just that it's poly-time vs non- poly-time, but it's actually, you know, very small program can emulate everything and provide exactly the same security, then I will definitely concede that point. But I have yet to see evidence. In fact, all I see the evidence I see, as these systems go live is that the surface area of things you need to emulate Proof of Work exactly.

Kevin:

So yes, first of all, thank you for the suggestion on Equinix. I will definitely look at that. Always looking for great investment. That's a great point. Um second no, absolutely. I don't think anybody has done a full on complexity theory study on the reduction of whatever the class of algorithms that Proof of Work encodes, and how they can be translated into, or rather, if Proof of Work encodes only super-exponentials set of algorithms, and we can't really compute them in any polynomial time on our machines, then that could be something to consider. But it's really not the case. I see no reason. That wouldn't be the case. especially because let me propose to you a design that it's not perfect. But it probably is, you know, ge nerally i n the right direction. Proof of Work, really what this process looks like is this pool of capital is just being generated out of thin air and that is being injected into the global ecosystem. And in fact, you can decouple proof, in Bitcoin, you can decouple mining from validation entirely. You can imagine Proof of Work, a variant of Bitcoin where there is one chain which is generating nothing but just new coins, is not violating any transactions. And then these new coins are simply being airdropped into this Proof of Stake system. and those people t hat are, you know, managing the o ther validation of t he system. So you can imagine, y ou going one step further where you make a Proof of Stake system, where there is one network which generates out of thin a ir, coins using Proof of Stake and then transitions them into the second network. These are two fragmented networks, such that if the first network ever capitulates and there are no coins being generated, then it does not prove any security issues, at least very, you know, obviously, at least to me right now as I'm sort of muddling my thoughts. To the second system, and the second system can continue operating even if the first one disappears, which effectively you know, looks more like something off the entire Bitcoin network, sort of going offline you know, rather mining going offline and, you know, the plug being pulled, yet the coin sort of still existing under the ownership of a group of individuals that are running a Proof of Stake system. So you know, this is very muddled, unclear fully spec design. But this seems to go in the right direction and it seems to mostly emulate what exactly Bitcoin is doing. So I don't see this, with a few kinks worked out, I don't see why this couldn't work. And I think the argument here is that, Oh, seems complex, therefore we shouldn't try it. That's not really a good argument. I mean, ultimately this is money, this thing. That's, by the way, one of the most awesome discussions that I would love to have, which is in regards to really what is money. and I've heard all kinds of arguments coming from all sides. you know, Proof of Work, people say that, you know, Bitcoin is valuable because there was energy expenditure that had to be put into it. Therefore you know, sort of the floor of Bitcoin necessarily must be the energy expenditure they went to it. except that, you know, very obviously this is not the case because market theory, which is what really dominates us nowadays is that, you know, the value of something is nothing except for what the market is willing to pay for it. So it does not matter what energy expenditure for Bitcoin is. It all, every, every single asset is really only as valuable as somebody else is willing to pay for it. So that's really all that it boils down to. So all those arguments on whether even Bitcoin is valuable kind of fall apart. So it doesn't matter how much energy expenditure has been put into the system. If miners decided to dump everything and crash the system and everybody flights out of Bitcoin, that system is now entirely valueless. And there is absolutely nothing that backs this up. this is a almost existential and highly cognitive dissonance inducing statement because people are like, Whoa, no, this thing is decentralized. It can never go down. But this is in fact exactly the case. If people lose faith in this, this i s done for, there is no reviving Bitcoin back. So it's really ultimately down to what people believe. you know, the value of the system is that's really what makes this valuable. That's what and then what kind of uses can, can, you know, that value be used for in B itcoin i s pretty limited in Ethereum. It's, it's much more broad. which is great. But ultimately it's really about what the market decides this can be used for. And t hat that's really what makes it money, not some exogenous, thermodynamic process that generates some new things out of thin air.

Richard:

Okay. So I think your last point about bitcoin and Proof of Work system seems to basically defeat the argument that burning of energy does not necessarily make something valuable. Okay. So, but how does that argument apply to POS' superiority or benefit over POW? Does that argument not apply to POS coins simultaneously?

Kevin:

100% it does the same exact argument. There is, the security of Proof of Stake system is really the best calculation you can make is at that time what the value of those coins is, because the market set the value of those coins, there isn't really any other inherent security to it. It's just the market decides is what people believe this to be a value that so no, no, Proof of Stake is absolutely not immune to those arguments.

Richard:

Okay. Okay. Would you like to respond to that by the way, Tarun, maybe not necessarily the last point, but any other points you mentioned?

Tarun:

Yeah. actually, I mean, I loved the last point and I'll explain a little history of how that's how I sort of got into this universe. But the one thing, I guess I, the only real initial comment is I wasn't necessarily saying that because it's complex. Oh, we should give up. I just think that because it's complex, you're not going to get like random person using it. You're only going to get like HFT people, right? Like at some level, the complexity that exists right now is like, you're only going to get kind of like the people who are smart enough to understand otherwise they're going to get screwed. And I think that that's just like fundamentally different than how Proof of Work is sold to the end user, which is plug in machine, hash. You know, it's like, it's just the cognitive, like, and I think it's great that like, you know, these systems exist because they're complicated. but I am just trying to point out that if we want to get to a world where like these things are used widely by everyone, then you may and may make these systems still seem a little too complicated. That being said, I mean this,"what is money" question? you know, I think I ended up leaving trading because I as trite as the story will sound, I was at Burning Man and I was trying to ask myself the question. That's why I had to add this, this"trite," right.

Kevin:

Nothing to be shameful of. It's fantastic.

Tarun:

And so I was at Burning Man and you know, I was talking with my friend a bunch about you know, like what does the definition of value like, which is sort of very closely related to this question of money because I have this really good friend of mine who I sort of actually sometimes help with some art projects. She did this kind of crazy art project in 2015& 16, where she drew random line graphs and in front of an audience. And then she, she went and, and went on like Charles Schwab and then she went and sort of manipulated the price of penny stocks to match the line graph. She just drew and then she auctioned off the paintings immediately and she sold a lot of them, like hundreds of thousands, or millions of dollars worth of this, which is, it's just basically market manipulation portraits. And so that's one of those things where, and afterwards she, after she'd sold everything, she asked everyone like, what does value, what does value mean in this entire event? And that was one of the things that kind of made me have this epiphany of like, okay, yeah, really money is definitely a large portion that's consensus, but then there is a large portion that's like just raw resource consumption, human life surviving biology type of stuff. And somehow I think during, during that, that that year at Burning Man, that was when it clicked for me that cryptocurrencies are the only thing that I've ever seen that has been able to directly convert one type of value to the other with, with like without exogenous kind of things happening where like you have this resource driven value, which is kind of this endogenous, you know, you know, you need it. So you're, you know, at a certain point you're willing to do anything to get it, versus the type of consensus based value, which is, you know, valuing art or valuing stock manipulation, scratches. And yeah, there's something, there's something innate to Proof of Work in my mind that connects those two things. It's like this resource that if we don't consume, we die. Which energy is not quite that, but very close to that nowadays. And we some have converted it to something that's like a pure consensus thing, and then maybe that Proof of Stake will be able to obviate it and make it fully consensus based only. But in my mind, there still needs to be some, some tiny connection to this sort of intrinsic biological value. but maybe that, maybe that may be, maybe that's false.

Kevin:

That's probably, that's probably one of the most interesting parts of this debate. and I got into Bitcoin God, 20, 2010, 2011, like really early on, precisely because o f thought of that process. You know, it was one of the few things that you could think of that was very seemingly elegant taking, you know some resource that was, you know, physically difficult, whatever to, to m anufacture and then converting it into, into money. The, you know, deeper you dive into that though, the more the sort of dream sort of fades away a little bit. unfortunately that's just not the case with Bitcoin. It's not a, you know, something like, like, you know, gold, which is truly a ph ysically s carce resource. And you know, it's just, money is j ust manipulated by the art process, which happens by just raw processes that happen in space, thermal, what is it called, nuclear fusion, whatever it may be. I'm not really an expert here. Tarun, you're the physicist. So you know, you know, this stuff. But the concept of, you know, transitioning something and all expensive into money directly and it basically being a one-way function seems like an elegant statement. it seems almost like one of those Eureka moments, like a clean mathematical structure that I have something of value. It transferred into, into money and it's one way you c annot move back. Therefore, the new money that was created out of this expensive resource must be valuable is truly an interesting premise, except that ultimately in Bitcoin, there is the last step, the last mile. And the last mile is everything that people keep missing, which is that last mile ultimately ends up being consensus. And consensus is really the last gatekeeper to really achieving this amazing dream of really having this thing which truly, you know, we have transplanted from a resource that was very expensive to produce into money directly. And it's one way it's that last mile is breaks down this, this entire argument. if we instead, for example, we had this mythical new Bitcoin which generated coins out of thin air, based on some very expensive process. And these coins did not require any form of global consensus. That would ultimately be a much stronger argument, because you now don't really rely on what others believe this to be worth anything. This thing is worth it because I can take this without questioning really anybody else, without really having to rely on anybody else and, and transact with it. That makes it valuable. But ultimately, even that, you know, utopian version of Bitcoin, even that still has sort of a last meter problem, which is that ultimately the receiving partner needs to believe and needs to agree, achieve consensus of this thing is valuable. You know, the US dollar is not valuable because we, you know, have general consensus over it being valuable. It's because it, there is a gigantic army of people in uniforms and in machine guns, tanks and jails backing this up, that this thing is valuable, and there is no question about it. And if you do not think this is valuable, you know, there is consequences to it. That's what makes US dollar valuable. There i s no such thing for Bitcoin. Ultimately it is on the r ecipient's view that this thing is valuable. And ultimately even further for the current Bitcoin implementation, it has these crazy amounts of risks that are at the last mile with a consensus, with a global consensus actually maintaining this global ledger. So you could have systemic risks like network partitioning, like people believing the network and hash power reducing therefore it being susceptible to attacks. Total reorg, total rewriting of transactions. There are so many intrinsic attacks here that are possible just because they're possible with any consensus protocol. It's not just Bitcoin, it's any Proof of Stake protocol will have the same problems. Ultimately, the only thing that matters here is this last mile, and that is what people believe this to be valuable. And prime example of this empirically is a Ripple-- useless, but people believe it t o be valuable. So it's valuable. IOTA, it doesn't even work. It's not even out there. It's literally had been shut down for a month, but it's a top cryptocurrency because people believe it to be valuable. Bitcoin has zero of the functionality of Ethereum, has the same energy expenditure, maybe slightly more than Ethereum. Yet it's much more valuable because people believe it to be valuable. It has a brand that's really all that drives, ultimately, the value of all these systems. It's really what people are willing to pay, what y ou're willing to pay for the product that I'm selling.

Tarun:

So I guess one, one thing I would add to that is you know, if we look at the types of things that people value basically on consensus only, and not in a needs-driven way, we basically end up finding that the, and-- when I say consensus only, I actually mean things like art or things where the costs of production is completely divorced from the actual final resulting price-- in the case of something like art we, we find that these, these consensus based pricing methods have the problem that they're only really good for pricing Veblen goods. So Veblen goods are goods where the price increases as demand or as, as a sort of, yeah, like it has the opposite of your normal behavior, right? So normally, you know, the price of something increases when the demand goes up or the supply goes down in a Veblen good, that's actually a weird thing where, where the price goes up as, as the sort of, you know, this cachet goes up,

Kevin:

Rather, the demand goes up as the as the price goes up.

Tarun:

Yeah. Sorry, sorry. So the demand going up as the price goes up is something that like always happens, seems to happen in these consensus space things, whereas things kind of end up reverting to a market price more clearly for normal goods that have traditional pricing, like, like the price of bread is super competitive. And it has sort of a reversion behavior where like once some participants start moving the price too far in one direction, there will always be someone who comes to undercut because there's very little stickiness to the brand in a lot of ways, versus these things like art where it's like so tied to this instantaneous brand that you need consensus on. So the question is, will cryptocurrencies ever be able to break out of being super useful for Veblen goods? Predominantly Veblen, cause at that point that's when it's kind of you really hit society.

Kevin:

Absolutely agreed. I think currently cryptocurrencies are effectively Veblen goods and you know, unlike other, I mean I see this breaking out the, that people demand Bitcoin for more than future price growth. If people are like, I need Bitcoin because it has a critical solution to a problem that I have, or I need Bitcoin because it will feed me and my family, then that until that, that's the case, then this is very hard to break out of a Veblen good.

Richard:

Got it. Great. The question from audience member Tanner Allard, this is actually a comment but it has sort of inspired the question. So he says if dominant means more popular, yeah, POS is more popular with you projects as a better mechanism for layer one. If dominant means more effective while preserving some ideal L-1 properties, then no. So you can respond to that. But the question that sort of got inspired from this, for me, is that new projects that are coming on board, why are they all POS, and very few are POW? And this is for both.

Tarun:

You don't need to reinvent the wheel when there's stickiness to the initial brands. And so, you know, we already know there's dominant POW competitors in privacy, smart contracts and just raw, sort of, value. So any new competitors getting into there would have sort of exponentially increasing hurdles, hurdle rate over the existing competitors. Whereas with Proof of Stake, it's still definitely a wide open field, especially if, you know, as Kevin said, there's like a use case, like, I don't know, you disintermediate some weird derivatives exchange that happens to trade, you know, like X billion dollars a year and maybe that like that. That's certainly probably more suited for Proof of Stake than Proof of Work. Unlike consortia, certainly.

Richard:

Well, but if you're talking about decentralized finance, is that a characteristic necessarily of POS protocols?

Tarun:

Yeah. So what I mean by that is actually there are like consortiums of exchanges that share settlement resources akin to kind of how DTCC is structured or, or, but you know, with derivatives exchanges it's very different clearing than it is with, with something like DTCC which is how equities are cleared. What I mean by that is like someone might make a Libra like consortium that like is just derivatives like derivative exchanges, smart contracting language and each of them can implement their local exchange as a smart contract, but they all have some stake in this consortium and that there are...

Richard:

Okay I understand now.

Tarun:

Yeah. So sorry, that was sort of maybe a, a complicated example. I do think an interesting place where Proof of Stake is getting even more stress testing is really in the DeFi land because DeFi contracts, their, their token values are really tied to true cash flows. Like their cash flows from lending, or cash flows from derivatives in the case of like synthetics. and people are voting on those cash flows. And like, I think we're going to have like the really fast pace Proof of Stake experiments in DeFi, and then the kind of slower ones in the layer-1's because there's so much bigger of a surface area. But yeah, I think it's just because there already are dominant players. It's kinda really hard to imagine beating Bitcoin at what Bitcoin does.

Richard:

Okay. Got it. Kevin, input from you and then we'll move on to concluding remarks.

Kevin:

I definitely agree. It's very hard to beat Bitcoin at brand. God knows people have tried and people are trying. but the other reason is you can't get to actually having th e, you know, if you want this to be used at hi gh p erformance and you want this to be used globally, Bitcoin can't actually do that. this is not a failure of Bitcoin alone. This is actually th e f ailure of all Proof of Work systems. I'm sure Tarun knows well, Proof of Work systems are these synchronous systems that have a bounded bandwidth, they have bounded performance r e quirements and properties. So they literally cannot scale. It's a fundamental limitation. So the best system that we h a ve k nown or b est mechanism to, to get out of that mode of synchronicity and this, this high overhead, that Proof of Work encumbers is through Proof of Stake. So it's a bit of a weird situation that Bitcoin is in. It's in this weird limbo state where it can't perform, it can't be used as day to day money. The best argument for it is a store of value, except that even as a store of value it can disappear at any day, hard to believe, but it could. It's not backed by force, is not backed by anything besides just the whims of a few miner s. So even that argument ultimately at the last mile kind of breaks down because people could just believe this not to be valued at anything. So consensus could break this you know, population consensus could completely destroy Bitcoin. So even as a store of value, it doesn't really make a whole lot of sense.

Richard:

Okay, great. Time for concluding remarks. So we will have Tarun go first. So just sort of synthetize your thoughts and then maybe mention a few things you picked up from Kevin along the way.

Tarun:

Yeah, so I think, you know, fundamentally I would say this, the notion of Proof of Work winning over Proof of Stake to some extent, you know, maybe you measure it on market cap, maybe measure that in total number of users, maybe measure that in non HODL'ing addresses, however you measure it. I just think that as its current state and with our current knowledge of the security model, it's extremely hard to imagine that within a year or within 10 years that, well maybe not 10 years, like within five years, that Proof of Stake will actually be able to be treated mentally by its users in a very similar way to the way Proof of Work is. And the other question is a question of approximation, will it ever really be able to get to the same level of capital efficiency in a short time span? and so I think those two kind of compounding factors make it a little bit hard to imagine that it will be able to make up for the lost time of kind of the branding and consensus and ledger state that the Proof of Work networks have that has kinda been deemed valuable. Now, I think Kevin and I agree in a lot of places, so it's not necessarily, I think it's more that we agree on or disagree on the notion of timescale and on the notion of you know, what, what it would look like for kind of a, a reasonable, usable Proof of Stake universe to, to occur. And I guess maybe I'm, because I think of everything from the lens of really trading, it just does not seem to me that we're kind of, we've kind of figured out, even from the market microstructure standpoint for like how these Proof of Stake coins should, should have a valuation model. There's not really a very clear model and not that there's a super clear one for Bitcoin, but at least you have a kind of, you've bounded the space by energy costs. So I leave you with the regards of, do you believe that first mover advantage will last forever, or do you believe that, you know, someone can really overtake kind of with the current competitors. And my belief is that the complexity will make it harder to scale to more users. And that could change, but just does not seem like it's quite solved. It doesn't feel like it's a solved problem.

Richard:

Okay. Great. Your turn for the closing statement, Kevin?

Kevin:

Sure. Yeah. So first and foremost, obviously Tarun is probably one of the people that I respect most in the space. Um but the arguments on on this being viewed from sort of oh, this is simple, tried and true, and we're reducing the space of our programmable functions down to just energy expenditure-- It seems reasonable, but ultimately I don't think it's a very strong argument. Things you know, that initially used to look very complex. People understood them more and more, and they became the new standard. but to concede, Bitcoin and Proof of Work have branding. And especially people outside of, this group of us in the crypto space, don't necessarily get the complexities of what Proof of Work means versus what Proof of Stake means. So they tend to rely on just, you know, branding really like what are people saying about something versus another? And obviously, Bitcoin has had a 10 year advantage on this, and a lot more stakeholders than Proof of Stake systems coming in. So there is a real existential threat to Proof of Stake, just not being taken seriously, or rather having to prove more than what Bitcoin had to prove. Bitcoin just had to prove that it was this new thing. whereas Proof of Stake has to prove not just that it's this new thing, but it also does work, and that it has a real value over Proof of Work. So people that are building Proof of Stake systems will need to be able to prove that these systems are actually going to be used for high performance, whatever high- performance needs, whatever it may be. And that you know, people are going to move wholesale to them. So as a much bigger roadblock for Proof of Stake systems to overcome, then obviously Bitcoin has already had to overcome Bitcoin and just had to really overcome the narrative and nothing else. So I think Proof of Work systems in Bitcoin really what Bitcoin is ultimately down to, it is a globally at this point globally agreed upon new assets that we can trade and speculate on. and this asset at this point has pretty well known properties as far as what it can actually be used for. Pretty great for speculation purposes. But much outside of that it seems to have limited exposure into beliefs about, you know, sort of value, which is obviously not the case because you need to really take a look at the full picture with clear eyes, I need to take a look at the last mile. That ultimately is something like gold is actually store of value because there is absolutely nothing besides just, you know, physical limitations to it that make it valuable. And not consensus necessarily even actually even with gold consensus really rules at the end truly speaking. But yeah, sort of rambling on thoughts here. You know, Bitcoin is going to be here for a long time is basically the point because it has proved something new and does not really need to prove much else. It just needs to survive for a very long time and I think it will survive for a very long time. Proof of Stake has a much bigger uphill battle. But I think the space for Proof of Stake is fascinatingly wide for how it can be designed, how it can be encoded, you know, what you can do with it. It's a really interesting space. And I think as we move forwards in the next you know, 10 years of the space, if we really want to go down the route of, you know really exploring and pushing the boundaries of what assets are and what and how they are, you know transacted and how they operate, then I think we'll really be seeing these more on Proof of Stake systems, where there's going to be more experimentation, more activity happening not on Proof of Work. You know, this is a store of value type systems only. So Ethereum is a great environment for this, but I think Ethereum at this point is also slightly outdated. but even though it's such a fantastic platform still, but ultimately, obviously all these guys can be overtaken. So I don't know. I think Proof of Work is here to stay, it's not going anywhere. But I would probably say Proof of Stake is, you know something to behold in the next 10 years.

Richard:

Okay, great. Kevin and Tarun, it's been an honor to have you come on the show and debate. So POW and POS has been an age old debate. It would be great to see how things unfold from here. So how can our listeners get in touch with you, starting with Tarun?

Tarun:

Twitter. It's my name, Tarun Chitra. Feel free to send me a message.

Richard:

And Kevin?

Kevin:

Yeah, same. Twitter at, also my name Kevin Sekniqi. Feel free to troll me and tell me how I'm wrong.

Richard:

Fantastic. All right, so listeners, we would love to hear from you and to have you join the debate via Twitter. Definitely vote in the post debate poll and also feel free to leave your comments. We look forward to seeing you in future episodes of the Blockchain Debate Podcast. Consensus optional, proof of thought required. Thank you guys!

Kevin:

Thanks so much. Thank you Tarun, this was awesome.

Tarun:

Thanks. I learned a lot.

Richard:

I would like to say thank you again to Kevin and Tarun. I suspect this will be an episode I'll revisit from time to time given its high information density. Indeed, the episode actually leaves me with more questions than prior to the recording. One interesting takeaway that I had, is that one weakness of POS systems is that, ironically popular DeFi applications threaten the security of the network by inducing token capital flight from security staking to DeFi staking. And another interesting takeaway is that liquid energy derivatives play an important role in helping POW miners achieve certain financial outcomes, whereas a similar mechanism seems largely absent in POS systems. So don't forget to vote in our post debate Twitter poll. This will be live for a few days after the release of this episode. Also, feel free to say hi or post your feedback for our show on Twitter. If you like the show, don't hesitate to give us five stars on iTunes or wherever you listen to this. And be sure to check out our other episodes with a variety of debate topics. Bitcoin store of value status, tokenization and smart contracts, DeFi, Bitcoin halvening, China's future in blockchain. Thanks for joining us on the debate today. I'm your host Richard Yan. And my Twitter is@gentso09. Our show's Twitter is@blockdebate. See you at our next debate!