ZKPodcast: Post-Devcon5 catch-up with Tarun, James & Georgios

Transcript

itto on the oil Chiti new duel [Music] welcome to zero knowledge a podcast where we explore the latest in blockchain technology and the decentralized web the show is hosted by me Anna and me Fredrik in this week's episode I catch up with Giorgio's konstantopoulou James Prestwich and Tarun Chitra we talk about their takeaways from DEFCON 5 what's on their mind and how they feel about the explosion in zero knowledge research [Music]

before we start we want to say thank you to this week's sponsor trail of bits trail of bits recently shared their review of this year's IAC our crypto conference on the trail of bits block they highlighted some of the major themes of the conference including the importance of getting basic cryptographic primitives right as they reported systems ranging from TLS servers and Bitcoin wallets to state-of-the-art secure multi-party computation protocols were broken when one small sub component was either chosen poorly or misused the blog also highlighted the ascendance of cryptographic tools for privacy-preserving computations such as zero knowledge proof secure multi-party computation and secure messaging systems for more on their thoughts from the

conference and what they see on the horizon in the industry check out a link to their blog in the show notes so thank you again trail of bits one last thing before we start the zero knowledge summit is coming up fast it's happening on October 26th in San Francisco it promises to be a day packed full of some of the best research in the zero knowledge space if you happen to be in the Bay Area for San Francisco blockchain week and want to check it out please do apply I'll share the link in the show notes I hope to see you there and now here's my conversation with Tarun James and George's what did you just say it's the first lines from a poem that are also the

alphabet oh cool so I'm sitting with James Georgio's and Tarun on this last day of Def Con welcome to the show guys hi it's good to be here nice to be back I want to start off with James and Tarun you've both been on the show before I'll actually put links to the shows that you guys were on in the show notes so anyone can find out what you've been working on maybe in more detail but why don't we do a quick intro and tell us what you've been up to since you are on the podcast let's start with you James so since last time I was on the podcast we've brought Bitcoin SPV to several more chains so we're now available in solidity going and ECMO script and we

have an upcoming rust implementation with wise and bindings we finally announced gbtc that we've been working on for almost a year now with Georgio's and other people and gbtc is the first project of our new trade group cross chain group which aims to work on applications across multiple chains who are we oh he is Summa Keith and a few other companies will probably be joining very shortly very cool Tarun what's new with you yeah we at gauntlet have been kind of growing and moving towards basically doing a lot more financial simulation for a theorem

smart contracts a little less on this proof of stake side I think there's been the there's kind of a little less demand on the kind of pure financial modeling side right now because right now everyone's focusing on engineering but in the defy world I think a lot of these contracts have got a lot of liquidity like compound and you know swap are starting to try to attract more sophisticated institutional investors and that they're kind of they're really needing stress tests and things of that nature so been focusing on that and have been thinking about a lot of new fine attacks on proof of state currencies

yeah cool lastly Georgia I want you to introduce yourself this is the first time you're on the zero knowledge podcast so why don't you tell people who you are what you're doing thank you for having me on uh so my name is George Oz currently I'm an independent consultant and researcher focused on layer two scalability solutions and the interoperability solutions always with security in mind first most of my work has been on the plasma side of things lately I've been working more and more on roll-up constructions like primarily the optimistic roll-up technique and lately I've been spending time with James on TBT C which is an amazing

protocol that you should all check and I've also been spending a lot of my time working on the IntelliJ protocol which is a protocol for interoperability that is agnostic with respect to their ledger of their settlement layer that you using and it's and like so you mentioned roll up here but I think of you very much as like also existing in the zero-knowledge community I try to keep up cool alright so now we're here on the last day of Def Con we're in Osaka that's where it was this year I want to hear from all of you what your thoughts were and I think maybe just start let's just do like an overall what was your overall thoughts of DEFCON 5 here in Osaka I think something I find

kind of interesting is that other large public blockchains are the biggest sponsors this year and there were a lot of other layer ones presenting at Def Con and I kind of got the feeling that the gaga days of late 2017 and early 2018 where people thought you could bootstrap a new community for each and every one of these many layer one chains has turned into wow there's a social proof you need from the theorem community if you're going to expand and we saw things like uh uh you know supporting eath we saw well like somewhat confusing launch of open Libre and you know it was interesting to watch

the the mix of of layer ones other layer ones the only other comment I'll add is it did seem a little weird that they got mainstage these other layer ones other like aetherium competitors were on the main stage but people working on it Syrian projects were in kind of off in different rooms and it was I feel like maybe it seemed a little weird to me a lot of my Def Con experience focused around eath - there were a lot of workshops and research breakouts and discussions around it I gave a presentation on cross chart communication you know generally the people working on eath to seem extremely

confident about it well the people who aren't working on it are generally more skeptical of it so I think that we're seeing many different implementation strategies for 8th - we're seeing them try to bake roll-ups into the shards themselves via execution environments we're seeing a lot of like very high level details of the later phases of eath to change very rapidly and I think it's perfectly reasonable to have uncertainty and healthy skepticism of it so that's kind of been what Devlin's been about for me is trying to figure out where eath 2 is going so for me

instead of being about where if 2 is going it's where layer 2 is going because we saw like the result of a very big effort they state channels initiative we're like 8 different projects merged and standardized on a specific way of like talking to each other on the protocol layer on there like networking layer and so on but currently like the design space is like changing in the sense that everybody literally everybody is going is what wants to do the roll-up thing and right now everyone's trying to figure out what's the optimal trade-off space because currently like they're all up seems to be the layer one

of layer twos you have the layer one the the public blockchain are you using have they're all up which just dumped data on and then you will be able to perhaps like plasma on top or channels on top and so on so a lot of my time it's figuring out it's been trying to figure out like what will they layer to stack look in the end will it be roll-up oh yeah it wouldn't be layer one top of it to roll up on top of it plasma on top of it channels on top of it in that ledger maybe like for everything speaking to each other how many layers is that we've we're way past layer - yes also every time someone says roll-up I can only think of fruit roll-ups so actually to that point one thing I think

is funny is the word roll-up in the last two months has dramatically grown in scope if you look at say III search or telegram or anywhere where people are talking about that stuff to the point that the word means almost anything you can you can imagine that has a data availability problem but there's zero knowledge roll-up there's an optimistic grow up and they have very different security and performance guarantees and yet people kind of you know if you go to some of these talks people say roll up for a little bro up but they don't like specify which one is like is it sort of replacing plasma or like if it it's like what where is roll up exactly is that because to me it's it's you know about sort of like an off chain or side chain

and being like brought back on so the difference between their roll up which is exactly the same literally exactly the same technique as the one that's going to be used for execution environments is that instead of you're just dumping raw data somewhere instead of doing putting everything off chain it's an on chain data protocol with off chain execution the optimistic roll-up uses fraud proofs let's take a roll up uses validity proofs the data that you put on chain is different in one case and in the other and the roll-up what it solves it doesn't solve the data availability problem but it kind of like

works around it by saying that we Longley advanced 8 if the data is put somewhere that we can all verify where it is and by solving the data availability problem quote-unquote you can have like arbitrary state transitions happen while in plasma or the data availability problem still exists you're very much limited in to what you can do in order to maintain safety of course data availability is very expensive you can think of it as one of the primary problems that public blockchain solve and that's why any kind of roll-up only gets you a small constant vector improvement in throughput one other thing I guess about

roll ups is the design space is still quite large in terms of there's a trade-off of time for safety it's like an compute there's sort of like the three axes and I really don't like these trial Emma's so I'm gonna just say there's three independent three somewhat dependent coordinates can we call it truce triangle I prefer simplex because that's you know there's a simplex and you can generalize that to more than three parameters but the III I think like those trade-offs are going to be very important to which ones get adopted and used and I think you're seeing a lot of experiments that

are engineering driven and the roll-up space versus purely theoretically driven like in other parts of the zero knowledge world so I'm pretty well optimistic about that okay we can totally come back to these topics but before that I want to hear a little bit more about what was your favorite talk I really liked a talk on stable coins which the name of the top was delivering spirals causing instability of stable coins and which I found to be very interesting because they actually do modeling and like science which we're missing a bit in the industry for which is like very much into like Tyrone's expect these like actual modeling on

which shows how certain actions of actors can affect like the results in the differ it can affect the defi ecosystem while currently everyone's like just putting money in the defi ecosystem and we don't know what's gonna happen when everything blows up actually Taran what do you think about like are you you must be paying a lot of attention to this did you see that talk - I've actually seen a react is that tucking in in New York but I I think there's a lot of these there a lot of these types of financial attacks that people are really worried about as the total contract values go up you know I think I think getting from the hundred million dollars in a contract to a

billion dollars in the contract scaling jump is going to take a lot more analysis of both these contracts and kind of these weird liquidation events that look similar to a lot of things from the financial crisis but sped up the deleveraging the interesting thing is there was recently a day where eise had like lot really large price drop and a lot of the services that are supposed to close or CDP for you when basically like when you're before you get liquidated they actually couldn't get get enough block space because to actually do the operation of closing your CDP for you and that's the closest

we've actually come and the interesting thing is there's just enough competition for block space and there's not so much block space that it's actually very hard to execute this type of attack but in the world where there's actually more throughput and scalability it becomes way easier to execute this type of attack so you know I I think I spent a bunch of time talking with people that why we haven't seen some of these live and it's it's it's actually it's it there's kind of this the limited box this kind of helps you in some ways so I guess I would add one other thing which is that I think there's going to be multiple fronts under which scale abilities is changing within the space one is financial scalability like can we really support more than X dollar

equivalent locked into a contract that interacts with the actual transaction scalability and we're going to see a lot more of these interactions leading to sort of malicious but very interesting to think about attacks and events so to rephrase what you said you mean that currently these attacks are not successful because an attacker is not like fast enough can cannot like get his transaction in fast enough but if we get to know if we had like a unit shop run over channels for example and where the attacker could do an instant like state update this could have been like much more dangerous yes one other thing actually speaking of the other proof of sake networks and I think you're starting to see a lot of them

trying to build in lending and stable climb primitives directly into the base consensus layer so you know I think the most famous example is sello who you've had on the show pretty recently but I think even the other proof of stake layer one projects I talk to a lot of them want to find ways for validators to actually make income streams that aren't just keeping up with inflation and some of those income streams involve letting you lend against or borrow against your locked stake token kind of the same way you borrow against your eath in maker and I've been hearing a lot of very interesting crazy ideas because I think people are starting to realize it's

gonna be hard for validators to make money if they are just keeping up with inflation and taking an ever-shrinking PI of the basis point spread on that you know there are some other revenue streams that are available to validators we call it minor extractable value and generally it's executing the kinds of attacks that we were talking about earlier validators are and miners are in a privileged position to execute any kind of attack they want for no gas instantly because they make the blocks for sure and I think the roll-up stuff will change the cost the price you know that sort of cost plus as as we said

with you know swap but are you saying like another way that validators can make money is just to corrupt the system sounds horrible I'm not corrupting the system they're performing valid actions which abuse badly designed mechanisms code is law this speaking of my favorite talk actually it was it was definitely fill Diane's talk on free will and it was something about how do you quantify extractable value in a smart contract and like what what's of what are the formalisms to bounce a like the total amount of value that any single validator can extract by basically having a bribe in contract or making cartels or different types of

behaviors and I think that we're gonna see a lot more of that in the proof of stake world and proof of work it's actually really hard to because the actual asset that is earning you it is paying for sort of fees is separate from the way of generating it by burning energy you don't have a lot of these kind of extractable events there's a much smaller surface area and the other aspect is that proof-of-work miners tend to have more avenues of optimization than stagers do so proof of work your primary expense is electricity and you can always search for cheaper

electricity and hardware and you can always search for better manufacturing processes you are likely to save more money doing that than you would by pursuing minor extractable value in many cases i want to shill for james cuz you know he thought he doesn't want to show his own article but he hasu and Brandon Brandon Curtis from radar wrote wrote a pretty cool article on this trying to quantify minor trackable value for a proof-of-work chains that definitely worth reading it's been in the works for months and months and we've rewritten it so many times and I'm super excited it's finally

out the door it just came out yesterday I think it's hard to tell with timezones hasu deserves most of the credit it's his model Brenda and I provided input and review mostly but but going back to the derivatives idea are these protocols actually they're building it into it and it's not it's not like you're doing wrapped so basically like I'm in cosmos where you have a lot of tokens already staked I've heard proposals I don't know if there's actually live where they'd actually create like wrapped tokens on another chain it's like promises it's like um io u--'s on what's been staked or something to be fair if you use maker die or use compound you are sort of have the same

thing because you convert your asset into the wrapped version of that asset the difference is the next generation smart contract platforms are a lot of them are going to have the the staking mechanism as a smart contract so right now the actual code that checks hey you transferred X etherium does it meet all the requirements do all the signatures are they valid is this account allowed to transfer to this account more or less that's all done in the client code separate from the smart contract side but implementing staking like the rules of staking in the smart contract makes it extremely easy to generate the derivatives in the same smart contra or

in another smart contract all right are you talking about kind of the tokens that decided to stay on ethereum then it sees like the ERC 20 type protocols like where it's like it still lives on a theorem or these also like the the layer 1 new completely new protocols they have all of their validation being done through smart contracts the rules of validation are done through smart contracts so for instance if I want to set the monetary policy of there's an inflation rate of this amount that's just executing the smart contract that's that computes what the current block reward is and then when validator is chosen by it say a VR F they they the VR F is executed as a smart contract - that

chooses a random address and then give them a block generally speaking newer proof of stake systems have much less of a division between contracts and consensus do you think that's a fair description yeah they're I think the idea is to make consensus a first class object so that it's implement it as a smart contract and and I think you know Facebook in particular with move I think one of the cool things about the language is that they force the smart contract developers to think about move only semantics so aetherium is scarce and the scarce object can't be copied it needs to be movable only making that very apparent to the

developer who's writing an app or a smart contract that having that as a first-class object makes it much easier to implement the other client code just so smart contract itself can you just really quickly give me a couple examples of protocols that are coming out that have that construction that you just mentioned where it's very very close hello Facebook so on uh probably it's - I actually no clue well it's very hard to say anything about eath - beyond phase zero the original like the onlythe - was that caspere would be not before it was called if - and that was up Casper FFG would have been a smart contract on eighth one yes yeah so this idea has been around for a while it's

just that now we're seeing like people actually like start to deploy it yep so going back to this taking derivatives a little bit the main thing to know about them is that they're literally unpreventable you can't actually force someone to lock up an asset because I can always make a side bet with someone else I can always go to Georgio's and ask for liquidity on my locked up asset the question is how difficult it is to do that and how much like slippage or premium I will pay in order to get liquidity you can't prevent people from trading things they own so we're kind of seeing some chains try out the Netflix

model of staking derivatives where instead of having these off chain pirate derivatives that are freely traded some chains are trying to provide and you know approved okayed ecosystem force taking derivatives on chain so that they don't have weird side markets interfering with their consensus process one of the ideas is that it'll just be so much easier to use the approved market that undesirable from a consensus security perspective instruments will be less common and that I would add one other thing which is that they're designed with the idea of the cost of putting an Oracle or some

other way of getting an external price feed into the underlying network has to be a lot more expensive than the on train version so you basically try to adjust the economics of these instruments both with the transaction fee level and that like kind of turnover level to kind of make it too expensive to have to interact with like an off chain centralized exchange this is cool a concrete example of this happening in the past is tessie's were traded on multiple different venues long before tezo's launched the same thing was happening also for ICO tokens yep people were buying tokens via smart contracts let's say I have like X Y Z ICO I would

send my ether to my smart contract the smart contract we would check how many tokens I would get from the ICO it would like mint like this many coins at that smart contract which were it like little and then this would be 101 redeemable for the tokens from the ICO cloud sell contract once they were liquid right and so because anybody can offer this and multiple major exchanges have done so in the past we should really expect to see each two staking derivatives available on the open market long before eath to has smart contracts or even transfers of assets what would be the relationship between them well

because you can always buy one staked eath two-for-one ether there must be a price cap you know it can't be worth more than one ether which implies that eath two will be worth less than ether buy you know whatever we value the staking revenue and if there is a way to go back from if to do with one would that change the price relation versus if it was only a one-way peg yes definitely yeah one one other way of thinking about this is if you make imagine you have two different prices in transaction fees for going from East one to e2 like one a

third it took one beacon ease and owed price going from one beacon eased some one ether and they're two separate prices you can think of the case where what's happening in stay zero looks like having one of the prices be infinite and the other one zero but in reality a smooth transition to this would basically be to make it very asymmetric so make it cost in transaction fees you know a hundred X what it costs for one direction relative to the other and then slowly bring that ratio down until it hits parity that would be a way to kind of smoothly go from having this price delta until you get it to converge right when you kind of are launching further phases and it's worth noting that it

will converge as eath to becomes more tightly integrated with eath one and you have to weigh asset transfers and things like that it's just likely to be very messy based on you know what we know about the market dynamics right now so I want to actually give a little bit of historical context to the this idea of having staking derivatives or like the sort of this theorem we're sort of sketching out that's claiming you must have derivatives at the moment you have a capital asset that you're trying to make scarce versus like a physical asset like energy that is naturally scarce and you don't have to sort of force scarcity

on it so in in kind of the normal financial world an interesting fact about us reserves there's much less money that's actually cash floating the US supply than there is other types of money supply so most US companies don't hold cash they actually have very little cash and one of the reasons that this has flipped since the financial crisis is that in during the financial crisis there are enough banks that failed and companies that had their money at banks without Deposit Insurance so in the u.s. every bank has to offer you deposit insurance up to $250,000 but

for companies which have bigger cash flows or have more cash sitting around they don't really that doesn't protect them enough they might have millions of dollars bank account so what happened was after that failure in order to sort of replicate Deposit Insurance this their Oh a market called the repurchase agreement market kind of grew a lot and what what how the repurchase agreement market works is actually very similar to a lot of these semi collateralized derivatives in crypto so I'll give you kind of like a high-level version of this so instead of having cash you go and buy bonds and bonds are considered

US Treasuries are considered the most cash like asset so you hold your cat so your Apple you get one hundred million dollars today you go buy a hundred million dollars of Treasuries and let's say tomorrow you have to build a new factory and you need to get a hundred million dollars in cash to pay Foxconn or something then what you have to do is you basically lend your bonds out for cash and then you go and buy the factory and then because your Apple and you have cash flow look your lenders trust you to pay it back in like a week once you know I don't know you sold some more iPhones and this system looks very similar to a lot of the on chain lending platforms

it's just that the pricing mechanism is very different other thing that just what you just described is very similar to what was being done with the C die in the art die and all these like died there were those but like in a very short this is like the iterations are incredibly fast what you're describing like you say it like it was far in the past this was like six months ago 75% of US dollars that exist in bank databases are in the repurchase agreement market so very little of the real of dollars that are outstanding in the world actually are not held in bonds that's a lot of dollars given that I have you three here one of the topics that I'd

really like to talk about is sort of the state of things since we're just wrapping up Devcon we can talk about the state of aetherium what's happening in the theorem right now in 2017 there was a very distinct thing happening culture moment whatever 2018 also had this sort of like it was crypto winter but there's a lot of positivity there's a lot of excitement there's a lot of building people seemed pretty pumped how do they feel this year what was this sort of sentiment this year I think that people are sobering up right now it's no longer a situation where like you raise a lot of money with just a white paper you need to have a product and you need to

find some product market fit if this thing exists in this industry beyond like regulatory arbitrage and securities fraud we're starting like finally the like scalability is like coming soon trademark in the sense that like things are starting to like get built around the scalable protocols Defy is definitely like the narrative that will stick around and I'm looking forward like to seeing how the competitors will actually like react to this because the competitors of aetherium are like very aetherium

accepts everybody like that's why the film culture is like we're very well known for that like come-as-you-are like we're gonna take you in like and the s tyrone said earlier we saw that with the sponsors and the mainstage talks by competitors however my what I'm I haven't figured out yet is like what will the competitors do I would expect they'd copy any portion of the defy ecosystem that finds product market fit but you guys think thing in defy has product market fit yet I think the interesting thing is that at least coming from the New York side of things I do hear institutions and more traditional hedge funds as opposed to

kind of crypto native funds being interested in a lot of the online lending products more as just a way to diversify some of their lending risk ironically in some ways I think the thing that has the best product market fit and this is I guess maybe somewhat controversial is tether which is the ultimate tool of traders and regulatory arbitrage the other thing is the Fiat backed stable coins have not had had any blacklisting so if you go and look at USD see the blacklist function has never been called that's why it's kind of funny that it's getting a went on compound so you can basically get a kyc

dollar lend it on compound to someone who's not KY seed and kind of get around the system so a lot of regulatory arbitrage stuff is working really well there's also a lot of kind of India and China focus companies that are trying to be like betterment or wealthfront so you know you put in some Fiat and they help you get into crypto and then they go and lend and get you some yield you know I think there's some of this like personal finance type of stuff that seems to actually be the narrative seems to be striking a tone with people who are very different than the crypto native users so that's just you know my at least from

the DFAC perspective what I want to talk about like to you James and you thoroughness um currently I think that there is for me at least there's a big question how do you build a business that makes money in this industry making money and they have like very following like constraints no token your business does not reduce to being an exchange so onboarding of boarding platforms or exchanges you and I swap which you could argue like has like some exchange like structure how do you generate revenue without margins going to zero fast and without like other than like consulting not we see revenue like I

have customers that use the product the product makes money you know it hasn't escaped my notice that all three of us work in you know services to other projects essentially sumo tends to work with layer one chains building interoperability solutions turn at gauntlet tends to work on you know agent analysis for other projects proof of stake in defy systems giorgos i know you've been involved as a consultant in a number of research on projects on layer one and layer two so like we've all kind of answered that question the same way is a were providing you know like non-recurring

engineering work for other teams and those other teams are all VC funded or token funded which usually means VC funded for now world just trying to make revenue where we can and I think we're all really searching for recurring revenue something that's a sustainable product and so based on where all our you know like businesses are at here I don't know that you're gonna get a solid answer for how to build that we're all still searching I think and I will say I guess the the businesses that I think staking has helped launch goal and seemed quite unbundled right now so note operation services like bytes and trails

who is on on the podcast earlier custodians especially as people seemingly want these NPC style custody instead of strictly multi-sig and broker-dealers so people who actually can provide sizable liquid D for holders I think you're going to see those three businesses integrate its you know from a traditional financial view those services are all sold by one vendor so like if you're a hedge fund you don't find a custodian who's separate from your prime broker who's separate from a person who's getting your bond coupons for you they're all the same and it's

the same service and I think someone will find that aggregation at that layer that's like slight in my mind slightly adjacent to exchanges it's it's exchange plus plus I think that you can tell by their moves they're all are moving into staking they're all operating nodes I think they're gonna sell developer node operation like in fira or alchemy I think you're definitely seeing that happen which is also going to just shrink the margins but at least there's like a minimum when you have a physical like hardware thing that you can kind of go to I'm trying to figure out how can

we grow the pie rather than redistribute wealth I actually have a question I'm responding to your question that's a question but you know I really remember something called an NFT that was supposed to be the pie growing universe for like online items digital items what what happened to that narrative it kind of I thought I thought ERC 20s were supposed to be the pie growing thing it wasn't a like rich State on a chain supposed to be the pie growing thing and now defy is supposed to be the pie growing thing and I guess we've gone through a lot of things that haven't grown the pie that have always been supposed to be that thing so I guess the answer to that question is you guys don't know the answer to that question

it's a thought experiment on that level I actually wanna I I want to talk about funding so I think your question was very much like how does a business make revenue but I think there are other ways that companies projects are funding themselves sometimes they've actually added funding directly into the protocol itself founder rewards block rewards you've companies that have developed protocol and they're probably also going to move into staking and validation if you look at like the cosmos team and who the validators are in cosmos they've at least found ways to maintain their own livelihood through that yeah I'm wondering what what are your thoughts on all of these kind of like new ways of funding I'm generally skeptical of

anything built into the protocol because it funds things it funds like people and people have control of those funds without the obligation to provide any value in exchange for them this is why you're like I am a fan of revenue focused businesses is because if people are buying your product that means you're providing value you take something that is a very objective like the protocol rewards currently for example in proof of work like a block gets mind they worth your summer if you add like a way for like dev funds like that get distributed based on some

multi-sig then you took something that also perfectly objective and you made it subjective based on that multisig and my whole idea for joining this industry was I can have like something I can observe like completely objectively like work is objective and proof of work is objective so like I want to avoid any form of subjectivity built baked in the protocol tying in to that a little bit the whole like thesis of blockchains is that incentives create results and that if there is a protocol sponsored you know extreme of money being given out that people will compete to get it and so this is why I don't really understand

the push towards governance and you know like protocol issued funds is if there is money being given out people will compete to get it one thing I will say about that that is um it's kind of weird and a lot of these new next generation networks is that there's not the homogeneity and node types and rewards as there isn't proof of work in a lot of staking networks there are many different types of validators so some validators are actually executing the state transitions some validators might get rewarded for providing data availability some especially in the zero-knowledge world will get rewarded for running proof generation like an

encoder right and now you have an multi-agent problem where you have to design the reward structure such that you have a revenue split that everyone is happy with and you now get back to this founders reward problem with this multiple node types so think about it that way that's that's interesting that like because I mean it the decision but I guess in this case the decision-making of the breakdown would have to happen or would likely happen before the launch or do you imagine it actually being altered after launch has happened and then there's some small group that can make those decision how much money is there to make by altering it after the launch has happened I think you're going to see

both of them I think even in the more per more permissioned law chains I suspect we will also see some some of these revenue sharing agreements changed quite dramatically as we've observed certainly with Libra and change to the Association members wow I didn't I haven't thought about that I hadn't thought about like the that connection between the sort of the issue with the subjectivity of block rewards and the distribution of that with this these emerging protocols and all the different roles that validators and the various types of validators will take since I have you here and this is actually going

to be episode 99 of the podcast which is kind of cool I really wanted to talk about the zero knowledge space I feel like you've at least all three of you have been observing it or acting in it working with it there's a lot going on right now so there has been a so called the Cambrian explosion of schemes in 2019 we called it it's snark temper it's knocked over like we'll see if we'll have that snack November I guess member it's like I don't know currently we've been using we've been using a protocol that has cryptography like Bitcoin and theorem all use cryptography

that's been well known for more than 20 years like even well standardized wide support were very well explored everything all the new protocols are the being released yes they might have amazing properties yes they might get you constant prover time constant very fired like low memory no tested setup or continued like never-ending setup however you're using cryptography from 2019 but only 30 or even less people have read let alone understand let alone understanding and being able to evaluate the proofs it's very like far from each other so I may be able to read the paper and kind of

get the gist of it explain it to somebody but me evaluating the security proofs of the paper is a completely different level and it's usually the authors that are going to be able to do that only and like the maybe their academic reviewers in the in some peer-reviewed conference the situation is that when these protocols get coded you can have you have another like problem because you have the perspec and you have the protocol you have the spec and you have the implementation each time you move from one to the other you reduce like there's even fewer people that can do it so like the people that can currently implement and deploy these protocols are very few and like I would still like be

very very careful with what I'm going to be using from all of this and let me take a moment to say that at the zk summit which is coming up in two weeks I will be doing a presentation on an app of the zero knowledge proving systems that like are being used today with the latest protocols which were released and the title is going to be a growth 16 is not dead because god 16 is like one of them like best protocol that we have to date it works very well it's been battle tested tested to a good extent and I think they're like we should be a bit more conservative than rather than like trying to reinvent new stuff at the same

time like why don't we give a little bit of context for the listeners like what kind of stuff is actually coming out what's exciting like I think what you've just described there is like this is the issue with a lot out but what what is it like what is coming out snarks are like the most widely known of these and there are a few different snark constructions dating back to 2013 or so maybe even a little earlier than that 20 2012 honey tonic arithmetic programs compilation Gentry thanks 98 the eighth gold busted what they'll be talking about but I I mean the first more practical thing was really the gentry paper and I hate being on podcasts with people who know more than me so they

only go back like 7 years in practice right and the first ones were incredibly inefficient so for example when Z Cash launched their snark prover took like minutes to run and the verifier was much more efficient but still is computationally very expensive snarks have an additional downside which is they need a trusted set up so there has to be some trusted group of people that go through a process and throw away the toxic waste and do the whole thing what we're seeing now is many many different systems that improve on specific aspects so like Groff 16 is just almost a

strictly better version or a strictly better construction of the same kinds of proofs we saw in 2012 but then you get into things like sonic and supersonic and plunk and all of these other names that I'm going to forget each of them are improving some specific aspect of it maybe making the trusted setup reusable for many different proofs maybe trying to remove the trusted setup maybe trying to have more efficient provers we're seeing kind of exploration of some trade-off space that we don't understand very well yet we don't know what is possible and how efficient or small or pretty we can make these so we're trying a lot of things and seeing

what has useful applications I've noticed in they like as as groups are trying to implement what's coming out what is actually happening is groups that aren't deep in the research see so much flurry of activity they're like this is exciting but we're gonna wait until the dust settles before we start really deploying our engineering team our engineering efforts towards these things because what if a brand new like underlying construction comes out in the next month or two which is very possible and it changes a lot of it that's at least what I've been hearing from some of the teams which is it's kind of like that the other side is the double-edged sword right are they

what if an attack gets found because the proofs in the paper were I will say one other thing especially in this the style of proof that's used for a lot of these constructions is that there's not quite perfectly comparable assumptions between the different models and actually knowing you know there's there's this sort of like performance first time for security verse a few other things simplex trade-off simplex triangle that's why it's this impact that it has a very center in all dimensions it's it's but there's basically we're also we're also forgetting that we're drawing all these trade-offs on the same surface

but there's actually different security assumptions and a lot of these proofs and a lot of them use different types of reductions and so it's not just reading a single paper and I think georgeous was sort of getting at this it's like you need to know the whole dependency graph of all the papers beforehand and understand when a certain assumption was relaxed or changed and that knowing that whole body of knowledge ends up being really hard for a single like a single person at a company to really be the expert on right especially because this is an extremely complex field and the work is so new in like ECDSA we're typically using assumptions that are 20

years old like discrete log problem stuff there's starks which are based on hashes and those are using assumptions that are like forty years old and battle tests a lot of like zero knowledge proof coming out today are using knowledge of exponents assumptions which are like five years old really and we have new vdf stuff using class groups which is maybe 10 years old we're seeing a lot of cryptography introduced that is based on poorly understood assumptions or assumptions that haven't been tested in practice or relaxed versions of older assumptions it's extremely difficult for anyone who's not a full-time academic to keep

up with and yet we still try I don't try all that hard I think at least personally I will say I there was a while when I was very excited by what I would call the the zero knowledge porn that was coming out every week and I really I really enjoyed reading these papers and and having discussions with people but the the the flood of the the Gilgamesh event the Cambrian explosion sort of has made it hard to even care anymore it reminds me of proof of stake in 2016 and 17 you kind of had first you had Algar and then get a couple of I get Snow White from Rafael paths and LuAnn she and you know there was just after

that there was this flood of provable proof of stake protocol papers I remember reading them and I was is amazing like and then by the time I read the seventh one I was like they're all actually basically the same with like a very tiny little you know difference because they're all chasing after ICMA and also the reason some papers that I've read they're like intentionally made very very very complex because complexity cells the sex cells of the cryptocurrency industry complexity says if you know if you've met Zack Union you know that he likes to say that all other proof of state systems are just tender mint with extra steps the the infamous

theorem of the tendonman reduction yes I will say that I suspect once we actually have simpler for a lot of the zero-knowledge work that's coming out right now we might be able to kind of connect this morass of different theories and assumptions and kind of get a good intuitive understanding this oftentimes happens in mathematics where a difficult result is proved and it's proved in a really complicated way and then someone finds a much simpler version and that each level of simpler proof you get much more intuition into why it works whereas the first really difficult thing was someone really straining and like kind of

climbing Mount Everest for the first time whereas each additional one is a better and better Sherpa guiding up the mountain that's nice well I really want to say thank you to all three of you for joining me on this podcast on our 99th episode is there anything else that you want to talk about before we sign off yeah I guess I had a question for basically everyone here which is you know I think we've seen a lot of interest in using Starks and snarks for Dex's but what are some other applications you know outside of consensus and X's that excite you especially with regard to the implementation because if you think

about it whichever application or layer or protocol implements this any of these rivers River hire systems they will be designing it around their organizational structure and like whatever their application actually is and so it's interesting to think about how the code that gets written will reflect the first user who's writing it so like what what are some of those applications you're excited about I'm really interested in how we can use zero knowledge proof sin cross train communication which is what I work on all the time and what I'm excited about so we've been talking a lot recently about this concept we call observability is what can a chain

practically know and part of this is like what is deterministic and publicly verifiable but the other part of it is what is cheap to verify cheap and that you can put it in a consensus process and so zero knowledge proof have this great property that they're publicly verifiable deterministic and a lot of them have constant time verification so they can be extremely cheap to verify and we can verify extremely large statements about reality and other systems in a very small amount of time cost so that might mean kotas recursive snark structure making

statements about the entire history of their chain and evaluating that from within some other chains smart contractor consensus process I'm a fan of dark pools but nobody wants to talk about it because it might be illegal to do but I'm a big fan of the design and I think that once you knowledge proofs like get deployed more we will see like like some kind of unis whap with a built in dart pool to be more robust to like when the market participant wants to add a bunch of liquidity to it so that like nobody can see that the market moved yeah that that is something I would like

to see at some point yeah I mean I think I'm I'm quite interested in the interoperability use case for zero knowledge proof there's like I've done a workshop where we with Chris goes from cosmos where he actually went through like all these different ways that's their knowledge type constructions can be used in interoperability not only as like clients but like in all these different ways it was really cool and then I mean but the topic that I wanted to talk about here at Def Con was about the idea of privacy and transparency and this balance that can be achieved through zero knowledge proof and that kind of takes the form of things like compliance or you know interactions with some regulators a very unpopular topic

in the in some circles but I think that zero knowledge proof in that context could be very very interesting that's a that's a very real-world example that I'm using here but thanks for that question should I have what you do you have anything I guess I'm in the the frame of Georgio's maybe not quite dark pools but some other kind of liquidity provision without leaking identity because in the normal markets that happens via speed not via brains and so it's kind of interesting to think about like what happens when you flip that dynamic that will just trading look like but again the it doesn't feel like that's like a growing the pie use case just you know

to go back to earlier it's just cool yeah okay so listen thanks again this has been really cool to have this conversation with the three of you I hope you'll I hope each of you will come back on the podcast maybe I can also get you guys all back at the same time let's try this again thank you very much for having a sauna tyrone and james are among molecular people i most respect in this industry so this was very nice and very insightful to me yeah thanks for having us and congratulations on 99 you're you know you're almost a big milestone I'm excited to see what you do for that episode well Frederick will definitely be on that one I'm sorry there Frederick isn't here and you know

sucker with us thanks for having us and uh this has been really fun it's nice to get a like roomful of friends together and talk about the state of things and where we think they're going I'm excited about it still so yeah next year alright and to our listeners thanks for listening