🎙 Tezos Co-Founder Arthur Breitman Looks Forward To No-hype Crypto Shining Through

This week on the Defiant Podcast we speak to Arthur Breitman, the co-founder of Tezos. Created together with his wife Kathleen way back in 2014, Tezos mainnet was launched in 2018 as a proof of stake chain with smart contracts, a combo that has now become the industry standard. They were very much ahead of their time.

Yet, Tezos has been slower to pick up than other Layer 1s such as Ethereum & Solana. We discuss its upcoming upgrades and roadmap, and how Arthur sees a deeper DeFi and NFT ecosystem evolving on the network. He argues that with the hype and vanity metrics of the bull run fading away, Tezos approach of what he considers is more thoughtful building, will shine through.

We also talk about how this current bear market compare to previous ones, as Arthur has seen a fair share of volatility during his time in the space. We also talk about governance, as on-chain governance is a key difference between Tezos and other Layer 1s. Arthur dives into Tezos experience so far and how that can apply to DAOs.

🎙Listen to the interview in this week’s podcast episode here:

📺 Watch the video here:

🙏 Thanking our podcast sponsors:

dYdX has launched a Grants program with funding allocated to open-source builders. Come help build the future of decentralized trading at dydxgrants.com

Nexo’s fundamentals-first model helps you secure your assets and grow your wealth. The company now provides an increased insurance on custodial assets of $775 million. Learn More!

Avalanche Subnets allow anyone build custom, app-specific blockchains - the future of Web3 scaling.

APWine - get your future yield today, speculate on APY evolution or fix APY rates. Discover the latest yield tokenization trend in Web3 with 'APwine'

👀 Only paid subscribers have access to the full interview transcript below.

Camila Russo:

Here we are with Arthur Breitman, co-founder of Tezos. One of the first alternative layer one chains. That's not my favorite term, let's say just plain. One of the first non Bitcoin layer ones. Arthur, welcome to the Defiant podcast. It's really exciting to have you here.

Arthur Breitman:

Thanks. I'm happy to be there.

CR:

Cool. So as I was saying, you are really one of the biggest crypto OGs. You started working on Tezos with your wife, Kathleen, back in 2014. This was about the same time that Vitalik was working on Ethereum. You then launched Tezos's mainnet in 2018 as a proof of stake chain with smart contracts which has become the de facto standard right now for layer one. So you were pretty ahead of times with Tezos. And as someone who has been deeply involved in crypto for as long as you have been through all the offs and downs of it, I am interested in your take on the current bear market. Let's start with just the current state of affairs. How does it compare to past downturns?

How does the current bear market compare to past downturns?

AB:

Yeah, on the one hand, I think everyone who's been in a space for a long time has seen boom and bust. And so you can say, oh,crypto has this buzz that happened in 2011, 2013, 2017. It's always the same. On the other hand, you have the approach [that] says, no, this time is different. So there are definitely some similarities in a sense that we saw a lot of rational exuberance in 2021 and it's calming down a lot. So we've seen that in the past. What's different is that I would say since the beginning of cryptocurrencies was [the] release of Bitcoin, we've had an easy monetary policy, you know, Bitcoin came out on the heels of the financial crisis. So there was a lot of quantitative easing and a lot of rate reduction.

And we've mostly had that for the life of cryptocurrencies. It was the Fed kind of blinked around 2018, 2019, and started cutting rates. But we haven't had a scenario like this, one where we're seeing both a risk of recession and eliminated inflation. So it's not just a crash of crypto asset prices. It's, it's a broad market crash, everything's going down, cash is being inflated away. Equities are going down, bonds are going down. So in that respect, it is I would say materially different from the previous cycle, but I don't think anyone knows exactly how it will end up playing out, or I'm not seeing anyone who claims to know for sure is lying or possibilities themselves.

CR:

So what's the impact of the fact that this crash comes with just macro headwinds. Does it mean that before it used to be a purely crypto isolated crash? So it was basically crypto coming down from too much hype and there was just this natural floor of okay, prices have corrected, there's some fundamentals here, people are building, and then it can start to rise again. But now there is just some added selling pressure because there's not as much demand for risk assets. And in this case is the crash going to be worse, deeper,longer? What do you think is the impact?

AB:

Well, I think there's a different hypothesis that needs to be tested. You know, I think it still remains uncertain why cryptocurrencies have value and there's definitely several good theories for it. One of it is that is treated as a store of value, the censorship persistent value. So people might hold Bitcoin or Tez or other cryptocurrencies because they can have direct control of it. It's not anyone's liability. It's a standard argument for store of value, digital gold, with of course the caveat. It doesn't have all the history of the store of value. And the fact that a lot of people who buy these crypto assets don't necessarily fit a profile of people who might actually really care about sovereignty in an currency. So you can patch that and say, well, people are speculating that people will need self sovereign, censorship-resistant stores of value.

There's been narratives about inflation hedges, which I think you're not doing very well in this scenario where we have elevated inflation and crashing crypto asset prices. Other narratives have been known that these are platform for applications and the value comes from the fees, which are going to be burned or distributed to proof of stake validators. But there's one theory, which is like, oh, you know what, this is a weird animal. And it's somehow just a manifestation of a very low interest rate environment when essentially nothing has yield. You might as well buy this crypto asset and it's not something that happens in a normal environment where rates are already made. I don't think that explanation is true. But it's not been fully tested because we haven't had long periods of elevated rates and cryptocurrencies in the past before. So it's a new regime that is yet to be explored.

CR:

Yeah. Will be interesting to see how cryptocurrencies can compete with rates. I mean, now people can just put their savings in the bank, get some sort of yield, which they didn't have before. And some of that money is flowing away from riskier assets.

AB:

Well, there was a take on Twitter that I thought was quite funny. You had this Simtel lab say, you don't know well, now that the rates are raising. This kind of Bored Apes are crashing and so on and so forth. And it was a quote Tweet by Neeraj who basically said, yes, now that I can earn 2% on my treasuries, I'm not gonna pay $200,000 for a monkey. I think it's a good way to reflect on the fact that, yeah, it's culture vs. Macro. The fact that people can earn a little bit more on the bond is not what's driving this. It's definitely something more complicated behind it,

CR:

For sure. Yeah. It's probably a combination of just many different things, just like crypto getting ahead of itself, macro. There's a war going on. There's just so much happening that can explain this crash.

AB:

Yeah. There's just too many inputs and too few outputs to really try to get a puzzle picture of what's happening.

CR:

Yeah. But I wanted to talk about some of the crypto specific drivers of this downturn. We've seen some gigantic failures in this recent crash. Obviously speaking about Terra, Celsius, Three Arrows Capital. It's, it's been so interesting to see how these companies and funds have unraveled. I think it's something pretty new that we didn't really get in the previous crypto bear markets because just the space wasn't as sophisticated back then.

AB:

We haven't had profiled blowups, you know, the blow of Mt.Gox.

CR:

Right. Mt.Gox. But yeah, but I guess this systemic dominoes falling with Three Arrows Capital affecting all these centralized platforms, just the complicated unraveling like this. Algorithm makes stablecoin. I don't know. It seems like the space got a lot more complex, so there was more room for failure. But I'd love to hear your thoughts on these crashes. What key takeaways are coming out from all of this?

Key takeaways from this crash

AB:

Well, I think the first takeaway is that the crash from Luna was highly predictable and predicted. I've talked about it many times of the fact that Luna was gonna crash. I've said it over and over. It didn't stop any of the large funds from taking exposure to Luna. And then it did crash. And I think the takeaway should be that fundamentals matter. We've been in a market that's so driven by hype and so driven by trend-following that in some sense, nothing matters. Every opinion is valid. And if something is trending and people are talking about it, then somehow that makes it safe. I think Do Kwon was very good at deal making. He got a lot of deals for himself into the crypto ecosystem causing himself up to other L1 chains.

And I think there need to be some repercussion for that. T's just, everything feels superficial in a sense like, oh, you do something with Luna, well, it doesn't matter no one cares that it doesn't actually work. So I think that needs to be the repercussion. People need to be a lot more careful and look in details at what they're doing, as opposed to trying to say, I'm going to buy this and then sell it to grid default because, hey, it's going up. It's a number go up mentality or up only mentality of the past year that I think is at fault. But what I don't understand is we've seen sophisticated players, sophisticated funds basically fault for Luna. And today, if you look in the past, you've had competitors in this space, in this industry like myself, who've been saying, hey, you know, Luna's not gonna work. Luna's broken. This is a problem. And people who said, oh, Luna is great. He's gonna bounce back. It's fantastic, even as Luna was failing and there's no reputational hits. And I think that's a problem. We're gonna keep failing this thing until people are held accountable for propagating these bullsh**.

CR:

What were some of the red flags or how could you tell with such certainty that Luna was gonna fail?

AB:

There were a few signs. First of all, it already had to be bailed out. So it happened in May of 2021. So it was not their first problem. The second thing was that they were promising this 20% interest rates, which wasn't in line with the borrowing rates. You know, the borrowing rates were high, right? The funding rates were high, but this was clearly not sustainable. And also they didn't have a mechanism in Luna to ate the rates as a function of the issuance of UST. So issuance of UST could go up and you would still get this very, very high rates. And it's not natural, essentially that you have that much issuance of the saying at this level of interest rates. So what naturally happens is that these are systems which work very, very well when they are in an expansion mode, right?

Because more people want it. But as soon as you have a contraction, it can't handle it. You have a contraction. So people start redeeming it against Luna token, which are minted, which are sold, which suppresses the price of Luna, which creates more pressure on the devices. It can depeg very, very rapidly as they shrink. And there's some sort of imaginate between where Luna collateral was so tied to the stablecoin that essentially if the stablecoin failed collateral would not maintain its value. So they were caught in this loop or big where as long as it grows, it works. And as soon as it stops [and] starts shrinking, it collapses immediately. And that's what used to happen.

CR:

Right. So the general lesson from Luna is to, it seems like pretty simple, but actually do research on the stuff that you're buying. If like you look closely at Luna's economic model you can come to that conclusion that in a downturn, when there's no demand for Luna itself, there will be nothing there to back the stablecoin. So that's not a sustainable system.

AB:

I would say that the lesson in general is that broad popularity is not a reliable indication of success. It's an important one of course, because we're talking about networks, we're talking about network effects, so popularity matters. But I think we were in a market where it was the only thing that would matter. And even if you had something that was financially and sound, it didn't matter so long as you were popular, so long as you were in a Z gasm, so long as it was buzz around it. That's basically what did this. It doesn't matter that Luna's fundamentally broken there's buzz. And that's what did it. The cascade failures, the businesses around it are quite interesting. I think the general theme of the past year is that there was a ton of leverage in the system. People really, really wanted to borrow against their coins, but there was not a good way for institutional money to come in.

It's easy now to look at the rates of 2021 and say, oh, well, the rates were representative of risk, right? The rates were so high because the risk was so high, but I don't think that's true. I think the main reason the rates were high is because it was a shortage of capital. People really wanted to borrow a lot to buy more crypto assets and most institutional lenders were not comfortable getting into the crypto market. So there was a shortage of capital. There was a very, very profitable trade over the past year in being basically a platform where you would match lender on one side to borrower on the other side. So people did this with defi protocols and let's set Luna aside because Luna is just like broken economics, but defi protocols, which are not based on broken economics, which are based simply on collateralized lending have fared actually very, very well in this contraction. It's simple. And it's almost amazing that companies like Celsius and BlockFi managed to blow up in this environment because this was an environment where you could make money, hands over fist with a very simple business model which was get a bunch of collaterals and then blend against it and then do margin calls.

CR:

So why do you think they had such spectacular failures? Is it that they were overly greedy and they were doing more with the collateral that they were speculating and putting their customer's money in illiquid investments, and so on?

AB:

I mean the advantage of using greed as an explanation is that everyone's greedy, right? So if you're saying they were greedy, well, who is not greedy but if we define greed as being blinded by the short term appeal of profit if you look at what happened, I think, and I don't know the full picture, and I'm sure there's some people who have a lot more knowledge about this primarily technologists, I come from a markets background. I was a quant in banking. So I do like thinking about this, but it's also not my main area of experience, and it's not the one I'm most knowledgeable with at the moment. But my understanding is that there were a few trades that did-- either Celsius or BlockFi had uncollateralized loan to 3AC, who themselves took heavy losses on Luna and a few other things. But I also think Celsius was also into Luna. So taking losses into Luna, taking losses from uncollateralized loans, into other entities. So it's not even margin failing or liquidity failings, not even liquidity crisis failing, it's just being uncollateralized. So I think those were triggers. The other trades besides Luna that were problematic. One was the GBTC discount rate. So basically you buy GBTC ETF, and you hope that the discount is gonna close either because it gets converted by, as an ETF by the SEC or SEC allows them to convert it, or somehow, the digtal currency group allows redemptions, but neither happened. And in fact, they have incentives to try to turn into an ETF because it would be huge for them, but barring that they don't really have incentive to a low for redemptions.

So it's a weird situation and the prices can divert for a long time. And the other one was staked ETH. So people were buying staked ETH from Lido which is staking derivative that's available for the Beacon Chain. And you can convert it into if and when this era merge happens after an staking period but the merge keeps being delayed and delayed and delayed. And all of a sudden, once you, once people start actually wanting their ETH back and not their staked ETH back you end up with a crisis where you have to sell into liquidity market because you can't redeem. In some sense, we've seen these failing trades before in finance. If you look at long term capital management, they had conversion trades between bonds that ought to be worth the same thing. The problem is the market can stay rational or illiquid longer, and you can stay solvent. And so if you leverage a lot into these illiquid trades, they can blow up in your face very easily.

CR:

So there, there were several very popular trades that didn't play out. Luna,staked ETH, and GBTC. And then obviously, I would say reckless,the idea of borrowing huge amount of money to 3AC uncollateralized. So all of those things came together to blow up these essentialized platforms.

AB:

I don't have special knowledge here. I just wanna say, I'm telling you stories that I'm mostly reading in the news that I'm reading in Twitter, and that kind of makes sense. And I'm not seeing it really a challenge. I could be wrong. I don't have any special knowledge in the books of Celsius of 3AC or anything like that. So I could be completely wrong, but at least as a story it's coherent.

CR:

Yeah. That's public information that has come out and one narrative I've seen that's very interesting is what you've already alluded to, comparing defi lending protocols that have to work with collateral because that's how these smart contracts are built, versus, centralized lending platforms that, are more opaque and have more discretionary ability to deploy capital and can take more risk that way. I just think it's really interesting that we've seen CeFi fail, while DeFi of course, like TVL has gone down, tokens are crashing, but everything is still working. People can withdraw, they can deposit, liquidations are happening in an orderly way.

AB:

Oh, actually, so it depends on which protocol you're talking about. If you're talking about Maker, MakerDAO does it well, they do liquidation auctions. They weren't doing it completely well back in March, 2020 when it crashed and they had all the auctions simultaneously on the chain, but they fixed it. And then you have some protocols which are quite naive about it like Soland. And then when Soland ended up having this huge position that was gonna be liquidated, they did this thing where there was a single vote that managed to take the collateral away. And I think that is not what DeFi should be about. But the real DeFi protocols did quite well through this.

CR:

So I'd love to talk about Tezos. Can you explain the basics. What's it about how is it different from Bitcoin and Ethereum?

How is Tezos different from Bitcoin and Ethereum

AB:

Sure. So Tezos is a blockchain decentralized blockchain and smart contract platform. The main difference with Bitcoin and Ethereum is unlike Bitcoin and unlike Ethereum based on proof of stake. So it means that instead of being secured by people hashing and spending a lot of energy, the consensus group is based on who owns a token. And we were one of the first proof of stake protocol to launch, not the first, people had done proof of stake before, but I think we were the first proof of tech protocol, where there was a natural economic incentive in a sense that if you cheated the protocol, you would actually lose your stake, which hasn't implemented before. So that's what a big aspect for us as proof of stake. A second aspect is that we have a strong governance model where the chain evolves by a vote of its stakers.

So if you stake Tezos, you can actually vote in how the protocol should evolve. And then that evolves automatically. So we don't need to use hard forks in order to coordinate changes. And I think that's quite important because if you coordinate with a hard fork, it can put people in position of dictating what the future looks like. Because in some sense, let's say you on Ethereum and the Ethereum Foundation announces a hard fork. If you disagree with it, it doesn't really matter because most people want to be on the same network. That's the primary thing. You don't want to be isolated on a network and you want to be on a network where USDC redemptions are gonna work and you want to be on a network where all of this real world assets have been tokenized and are gonna keep working and you wanna be on a network where infrastructure is built.

So in some sense, being a shelling point, a focal point, in terms of who did you look for to know what's canonical and not canonical can give you a ton of power. Even if in theory, people are free to follow the hard fork or not free to follow hard fork. I would say a lot of that freedom is losery. What you have with a governance model is that it doesn't, you know, you're not penalized for voting against it. You can vote whatever you want so you can express a preference, but then you still know that there's gonna be a decision and you'll be on the same network as everyone else. So it's not economically costly for you to express your preference. And I think that's quite an important feature, but as a side effect, it's not just about philosophy and principles in practice. Building these governance system has given us a system that's very easy to upgrade. And we've done about 10 upgrades in 10 years of existence of Tezos with 10 just taking place at the end of June, right before we turn four.

A third aspect of Tezos is that we took a different approaches from Etheruem in terms of virtual machine. So Ethereum has a very low level virtual machine, the EVM, I think part of the motto of Ethereum is that Ethereum has no features. Iin some sense you should build everything you can from the VM. So it's like very, very low level, build everything on top of that, which has a lot of advantages. But one is advantage is that you can't really introspect easily into contracts. It means that when you look at the contract on a chain, it's not always obvious what it does. You have to rely very, very strongly on compilers, whereas with Tezos, we give people the option to either use compilers with high level languages or to even directly write very, very specific assembly for the chain. So it gives you a little more flexibility and it makes it easier to write, say smart contracts.

CR:

Is that thanks to using another virtual machine or is it because of the programming language you use? What's the feature that gives it that flexibility?

AB:

Yeah. I think it's the virtual machine. The virtual machine itself. It's a cross between Bitcoin script. Imagine if you wanted to take Bitcoin script and you said, I want Bitcoin straight, but I would like Bitcoin script to be turned complete and statically typed and a functional programming language. So you'd get something like a virtual machine. There's some confusions though. Some people imagine sometimes that they have to use a virtual machine. And in fact, they don't. You can write Bison and you can write JavaScript and run that on Tezos if you want to, but you can also directly write into the virtual machine, the fact that it's high level also gives us tool that can target the virtual machine. So if you're building a debugger, you don't have to build debugger for the high level language.

You build a debugger for the VM, but since the VM is very high level, very interpretable, you have a better view of what the contract is doing. I don't think right now, people care too much about the contract safety. Based on the market, we've seen a lot of defi hacks with hundreds of millions of dollars being lost and people don't seem to care all that much. I think if and when there's a wider adoption of cryptocurrency and more important use cases, people will start caring more.

CR:

I think people, people are starting to care more after the billions of dollars that have been lost in hacks of the past few months. I really hope the market is caring more about safety. But it's interesting. So what about the Tezos' virtual machine makes it better to write these smart contracts?

AB:

I don't think the visual machine is inherently safer. I think it makes it easier to do safe things. It's a set of things. So I'll give you one simple thing, integers are naturally inbounded so you don't run the risk of your indigenous looping around in the other direction. Some programming languages do that. So Python integers are inbounded by default. There's very, very small overhead in terms of computational time in dealing with inbounded and x?.But it's a nice thing that makes it also makes it easier to prove things because it's easier to prove things about in into than it is to prove things about surgery in to be integers, for example. So small thing, but useful thing. You have high level primitive, so you can have concept of maps and sets. You have concept of functions, all of that at the visual machine level.

So if you want to write a mathematical proof about your program, it's a lot easier to write it about this type of assembly than it is about some type of assembly like ZVM, or like <inaudible> because it's almost in a mathematical language, so it's easier to translate into medical proof. So that's one advantage it's also means that's when you write compilers, because you have these high level notions inside the VM, you have, I would say somewhat natural translation between the language being compiled and the VM. So the compilers are also simpler to write and the compilation is more obviously correct, and you can look at the result of the competition and see if it actually match the code you written.

CR:

Okay. So just generalizing, would you say that, because the Tezos virtual machine works at a higher level than the, the theam VM that makes writing code on the Tezos virtual machine easier, and because it's easier to write code, it's easier to check it and find bugs that way.

AB:

So first of all, most people who write contract do not write for the VM directly. Some people do, if you want super optimized code and you can do it manually, it's an option that's afforded to you. But it's not necessary. Most people don't do that. Most people use languages like Script or Bison to do it. I would say more useful for the tooling. So building debuggers, building testers, or building reliable compilers is I think where the main advantage is.

CR:

Okay. Great. And so it's interesting that you're saying that you can use any programming language to write applications on Tezos. I guess that's, that's different from Ethereum where you need to like write them on Solidity.

AB:

Oh, so when I say any, I think, there's a few programming, languages are very popular, smart bug is one GS legal is another one. Archetype. You shouldn't take any programming language to compile it to Tezos. You could do the same thing in principle. You could take any language, right. And compile that to EVM, but you would still need to write a compiler. And the compilation would not necessarily be efficient, but, I would say three or four, very popular, smart contract programing, languages in Tezeos are compiled to ?.

CR:

I wanted to go back to talking about governance because it's become such an important feature or topic now with DAOs gaining steam and we've seen the difficulties of having on chain governance and just like coordinating humans around decisions. So, you know, I, I'd love to hear more about your experience on how that's worked for Tezos, because all the problems that have come from coordinating DAOs, I've had conversations with people who are just like advocating for, e simpler the governance, the better. They're arguing for minimized governance. But I guess in Tezos, that's the opposite, right? It's like everything is on chain or what's the right way to understand it?

Minimized Governance

AB:

I completely agree. I am also a governance minimizer. I think you should have the minimum amount of governance necessary. But no less. And I think that a lot of don't have enough governance or rather they just don't have the right type of governance. Whenever you're gonna consider technological upgrade, you're gonna have a governance process, whether you want it or not. And that governance can look like hard fork based governance, but that's a governance process, or it can look like on chain governance. And I think on chain governance is a better way of making these decisions hard fork governance. The idea is say, oh, no governance. You know, we just, you know, we just run two different versions of the code and you can run whichever one you want.

But like I said, this is an losery choice because you still want to follow the pack. You want to follow the majority of the ecosystem. So it is a governance procedure. The governance is that people propose forks, and then you play correlation game where you try to predict. It's not about which fork you prefer, it's which fork you think is going to be the most legitimate. So it's an implicit unwritten governance mechanism, which is about cloud, which is about perceived legitimacy a lot more than it is about technical merit. Now, of course, if someone does a fork, that's horrible, no matter how much cloud the legitimacy is gonna have, maybe that won't work, right? If you come in and you say,oh, you know, this is the official fork, but somehow it burns everyone's balances.

And it breaks every contract, people who say, well, you know, I don't care how legitimate it sounds. I'm gonna stick with the old version, but within some reasonable bounds, it gives a ton of people, a power to whoever has this percentage legitimacy. And that's not the case at all. When you have untrained voting coin voting, you don't have this phenomenon. So I approve this form of governance, and if we could get rid of it, I would absolutely do that. And I think you should get rid of governance at some point in the future, when there's basically almost no more technological progress to be made, if that's the case, if you don't really need to upgrade anymore, then maybe you give up on governance. But for a while, as a progress in the space continues as progress in research and development continues, it's important to have this mechanism.

CR:

Couldn't the same thing happen with on chain voting, with what you're describing with hard forks. Like if you're seeing that all the vote is going to one option and all the biggest stakers and all the more legitimate, people with cloud, are voting in one direction, it can create the same effect.

AB:

I mean, if they're all voting in the same direction, right? First of all, you don't have to vote in the same direction. If they all vote in the same direction and you vote in a different direction, you're not gonna end up on your own fork. They'll win. They're gonna win the votes, but you still be with everyone else. So that's the biggest difference. With a hard fork it's possible for a hard fork that no one wants to still win. Because even though no one wants it, everyone thinks other people are gonna run this version and therefore they have to run this version. So you can have 0% support for a hard fork and a hard fork is still successful. Whereas that's not possible with on chain voting.

CR:

I see. So in a hard fork, the argument is that participants are more compelled to go to one direction because otherwise they'll be out of the chain. The consequences is much greater than just voting one way or another. Yeah. And what exactly do stakers vote on Tezos?

How stakers vote on Tezos

AB:

So the stakers, which you call bakers on a Tezos network. They make little breads, which are the blocks. They vote on protocol proposals, so anyone can come in and make a proposal. And a proposal is a hash of the protocol code. So you say, I wanna replace the protocol code with this new version of the protocol, which could be anything. So the hash is proposed and people can vote for it or against it. The process takes about two months and a half from proposal to adoption. And there's three phase of voting. In the first phase. You have an approval voting phase where people basically upvote protocol proposals, and it could be several at the same time. So you vote the ones you like, you don't vote the ones you don't like. At the end of this period, we take one proposal, the one that has the most upvotes, and we set it for a vote. So that's a binary yes or no vote. And that one has a required quorum around 60%, and the required approval of about 80%. And, we meet those quorums. So regularly you have quorums of 60% for the vote. It's not like, 75 protocols have governance and you'll see less than 1% of participation. We're at over 60% participation on a lot of these votes. And it requires an 80%, super majority. So it's very conservative. It's biased towards not crazy changes. If you piece off 20% of the network, it's not gonna go through. After that, you have a quiet period where there's no votes. People think about what they voted for. And then there's a second confirmation vote. And after that happens, there's another period where you wait so that everyone can get ready for the activation. And then it activates. And from start to finish, this is about two months and a half.

CR:

Interesting. And for the confirmation vote is 80% also required.

AB:

Yes.

CR:

Okay. And then after that two and a half month period, then there's another round of proposals that are evaluated. And so on. The process just keeps on repeating.

AB:

Exactly.

CR:

And anyone can propose,

AB:

Anyone can propose, we even had anonymous contributors.

CR:

But do you have to be a baker to propose?

AB:

Yeah, I think so. Actually you need to be a baker to propose. Or, you need to know a baker who will put in, the proposal for you. There's 400 bakers. If you write a proposal, you'll find a baker to put it in, in the same way that if you want to put a transaction anyway, you need to have a baker included in some form of another.

CR:

That's really interesting. And there's 400 bakers.

AB:

Around that. Yeah. It's hard to know because of course, you could have a baker used to different addresses, but there's no incentive for bakers to split their account. We make sure that we create no such incentive. It's very tempting for protocols to put their numbers by saying a staker cannot have more than this amount of stake. And then of course the big stakers will just split their stake and you can build boast a very large number, but in practice, it's a small number. I think it is ridiculous as how the Beacon Chain on Ethereum says, you're a validator when you have 32 ETHs. So it's like, oh, we have millions of validators. No, no, you have millions of packets of suits, 32 ETHs. And that should not be called a validator. A validator is one machine that validates. Yeah. That's a really weird terminology, which confuses a lot of people. Because we would say, oh, you have only 400 bakers on Tezos,we have hundreds of thousands on Ethereum? I know what they say it because they say, well, they could have different keys and so forth, but I think it's still a misleading number.

CR:

Yeah. So 400 baker with how much at stake?

AB:

So the stakers by themselves need two things to produce blocks. On the one hand they need to put up 10% of the collateral. So 10% of Tez at all time on the network are basically staked. And they also need to attract delegation. So essentially what gives you the right to create a block is all linked. If you own Tez at any point, sometimes you get the right to create a block. And that right is exercisable by whoever you delegate to. So if you delegate to yourself, you have to exercise the right. If you delegate to someone else, they can exercise the right, but they need to have a stick to exercise the right. So the right to create a block incomes on people who own tests, but the exercise of that right, is whoever it delegates to, and who needs a bond. And about 80% of the network, 80% of the Tez is delegated to a baker and the bakers by themselves stake about 10% of the money mass of Tezos.

CR:

So to be a baker, you need to stake 10% of Tez, but that's 10% of what, the amount of Tez that you own?

AB:

No. So I would say collectively. Collectively bakers will have about 10% of all Tez circulation at stake.

CR:

Oh. Okay. How is that enforced?

AB:

Mathematically through the protocol, if you want to exercise your block creation rights, you're gonna need to have the stake, otherwise you won't receive the block creation rights.

CR:

Right. Okay. So there's no minimum of how muc hTez a baker needs to put up. In the case of Ethereum it's 32.

AB:

No, there's a minimum. So the minimum is around 6,000 Tez..

CR:

Okay.

AB:

And it's mostly here for historical reasons. The number has been lowered in the past and we can keep lowering. But essentially it makes things more efficient to have a limit on it. And unlike ETH we do sortation of the bakers. So, and basically every block, you have a random sample of the bakers, which are selected to sign a block. So we don't need to limit the number of evaluators. We can have an unlimited number of validators. What the 6,000 does is that it tells you that your sample still has a lot stake.

CR:

Okay. Got it. Is 400 bakers to you decentralized enough? Is there a goal to increase the number of bakers? What's the path on that?

AB:

Yeah. I don't think decentralization is a number of bakers necessarily. I think decentralization is more about power in some sense, what powers do the bakers have. Because first of all, you could have, you have one baker with 98% of all the stake, and then a thousand baker was 2%. That doesn't mean that much. If all your bakers collude with each other, it also doesn't mean all that much. So first and foremost, the thing that's even more important than decentralization is minimize the power of validators. That's the thing you want. You don't want them to be able to harm the network. And that I would say is number one. So even though there's on chain governance in Tezos, somehow it has very, very stringent thresholds.

But even if they did something crazy, it takes place over two months and a half. You have plenty of time as a last resort to hard fork, if something crazy is going on, they're not really harming the network. So basically the work that validators can do is they can double spend, right? They can do reorganization, they can double spend. But the nice thing, which proof of stake is that you can detect that and you can levy the extraordinary financial penalties on the people who did this. And then they're excluded from the network. So proof of stake is very, very good at discouraging this type of behavior, but also as being self healing. If and when it does happen, there's a temptation, I would say to put in more task on top of what the validators do, but the more you put to more you increase their power. Most things on a blockchain do not need an honest majority. Consensus does. Executing transaction doesn't. Validating the transactions are correct does not require that you are in a full node. You can already validate that. So creating value blocks does not depend on the majority. It's really just ordering transactions and maintaining a single head for your consensus and you don't want have more than that. Now, once you've minimized all that, I think the question that are these people going to collude and start producing competing forks. I don't think you could say there's a number that makes this. I don't think like with 400 or with 100 or with 4,000, one's gonna be more like this than the other. The other thing I haven't talked about is censorship. So that's where I think the number of independent party is more important and there's two level of censorship.

Two Levels of Censorship

AB:

One level of censorship is I am a block validator and I refuse to include certain type of transaction, but I don't mind if other people do. So in that case, having a diversity of validators is quite important because even if your transaction doesn't get included by Coinbase baker, who decided they don't want certain types of transaction, it will be by another baker. And that's quite important. There's another form of censorships which is more OUS, which is I don't want to include this transaction, but I also want to block other people from including this transaction. So if you have a majority who wants to censor, or even two third who want to censor, they can not only censor the transactions, but censor other blocks who try to include those transactions.

So that's another power that the validators are. You can detect that to some extent. After a while you can do that. Transactions are not getting included. And there's an interesting proposal to try to fight censorship by validators. But I would say another one that matters was the number of validators. So is four hundreds bakers enough? I think so. I don't think tomorrow the network would be much safer if it had another 400 bakers. I don't think it would be much less safe if it had a hundred less. I think it's a good number. I surely wouldn't be happy if it had three or two that would not be good, but we are hitting the point of diminishing returns. And sometimes I'll often hear from the community how we get more people involved in baking, how do we get more bakers? We need more than 400. And I say, look, it's all good. The more, the better, obviously, but if you're looking at priorities, it's a lot more important, for example, to have fast execution of smart contracts or increasing throughputs than it is to somehow double the number of validators.

CR:

Got it. It's about kind of reducing the power of bakers. It's a great point. So speaking of that can we go over your latest upgrade, which is about increasing throughput and improving scalability on, Tezos. I saw this was announced recently. So yeah, I would love to know what this is about.

Latest Tezos scalability upgrade

AB:

So we just upgraded to Jakarta. So Tezos upgrades historically have been named after cities, especially cities that have been around for a while. So we had Athens, Babylon, Carthage, Dephi, Edo, Florence, Granada, Hangzhou and Jakarta. And so in Jakarta, we introduce optimistic rollups on the chain. And historically optimistic rollups on most chains are implemented as a smart contract because you can do that. But instead in Jakarta, you have optimistic rollup implement directly as part of the protocol. So called and enshrined roll up. So they have a special status in the protocol. There's direct support in the protocol, and that buys you a few things. Other thing it buys you is more efficient. It's easier to write code for the protocol than it's to write code for a smart contract. So development time is lower.

And the fact that it's enshrined also give guarantees to user it's not just going to be abundant because whoever was, working on a smart contract, decided to quit the project it's part of the whole Tezos project. If there's a censorship attack against it, then forking the entire chain as a result is a possibility, right? Because imagine tomorrow you have a censorship attack on Arbitrum. Arbitrum Is an optimistic rollup on Ethereum. One attack against it is you have an invalid commitment. Someone says I own all of th,assets inside the cell too. And I withdraw all them to myself. You do that. And somehow the validators of network collude to prevent anyone from challenging this assertion, which is false. So that's a censorship attack on rollups。 the question you would have is that is it worth forking Ethereum for this.on the one hand, clearly this is an attack by the validator set, but on the other hand, as they're attacking, they're not attacking the protocol itself, they're attacking one user of the protocol.

So you have this kind of ambiguity. And I think socially having something and try and removes the ambiguity. And it's like, no, it's clearly part of the protocol. It lets us do different saying the gas cost did. So there's a lot of benefits in having enshrined rollups and we're starting with one, which is fairly simple. It only does transactions. So it does very compressed transaction with aggregated signatures. So, you get down to just a few, like 10, 20 buys per transaction only. So you can do a ton of transaction on this rollup, which increases throughput of Tezos chain. And it's preliminary work for the next upgrade, which includes general rollups who can do a sort of smart contract. And these ones are based on a Wasmer kernel.

So it means that anything you can compile to wasm, you can run as a rope, which is also the approach that Arbitrum is taking. So we'll be able to take, go Ethereum for example, compile that to wasm and that's the rollup. You'll able to take some of the substrate chain that they built chain of port yet they've been built inspir chain of PDOT compiled, you know, that already compiles to wasm that's really as a roll up. So we'll be able to have a lot of different blockchain architecture run as optimistic rollup on the Tezos chain. So that's exciting. That's coming in the next proposal. And then after that, the nice thing with optimistic rollup is that you get to increase your super because you paralyze execution, right? So transaction that would normally happen by being executed by every node in the system now can be offloaded to a single honest node that follows the roll up. So that's one benefit. The second benefit is because you don't rely on an honest majority anymore, you can increase the hardware requirement for computing transaction without hurting decentralization. Because if I tell you, you need this huge computer to validate the chain to run a full node, you're gonna have very, very few people run a full node, and that's gonna hurt you. You need a lot of people validate your chain running and having a full node. You need an honest majority of lock producers with a roll up it's done on this majority. You need one honest party, a single honest party. So when you need a single honest party, it makes a lot more sense to increase a computational demand on that party.

So you have both this vertical and horizontal aspect to scaling with rollups, but you still have to include every transaction on mainstream. So every transaction on mainstream still has to be included. So your bottleneck becomes bandwidth, essentially, your validators of your network. They're no longer computing. All these transactions are coming, but they still need to download them and include them into block. And that's where that availability sampling comes in. And that availability sampling essentially says, instead of downloading the entire block validators download a portion of an error correcting code of that block. And once you do that, basically, I think you get to the point where you actually scale in a sense that the more nodes in your system, the more processing power you have.

CR:

Nice. So when do you expect the second step? The first step was already approved with transactions. Next step with general computing the smart contracts When is that?

AB:

So that's coming up in the next proposal, which I think should be proposed within a week or so, maybe a week and a half. But the one's being proposed now, its proposal will be part of the protocol and part of Testnet, but it will be deactivated by default on a protocol. And this is to give more time for us to help tool builders actually integrate with this new protocol to build some of those products to have more x? If I look at it. So if it's approved and voted and has a positive vote, then in three months, it would start kicking in, but wouldn't be active. It would take another two months and a half for it to be active. And then you would probably not. So L is the next one,M would be the one after that. And I don't think M will have that diversability layers will be in Testnet, but probably not in M. So I would say in N, that puts us basically six, nine months from now for activation of massive throughput, essentially,general optimistic roll up and adaptability sampling.

CR:

So Tezos as we mentioned, has been this longstanding proof of stake chain. But just looking at evolution of the defi ecosystem, it's been slower to pick up than some of the other layer ones, like Solana, Avalanche really blew up last year. Ethereum, obviously the first mover with this. So comparing TVL, Tezos defi has been a lot lower than on Ethereum, for example. I see 30 million on DefiLlama versus 30 billion for Ethereum. Are these upgrades going to make way for a bigger defi ecosystem or I've also seen Tezos has been focused on NFTs, maybe this is the direction that you wanna go? What's the direction for applications on Tezos in the next few months?

The direction for applications on Tezos

AB:

Yeah, there's definitely a lot of interesting things to build in defi and I've mentioned Maker several times. I'm big fan of Maker and there's definitely interesting different protocols. A lot of this thing we've seen and a lot of I think TVL has become a key metric in a lot of cases. And what we saw in the case of Avalanche for example, is, you know, people going to project and say, hey, we'll pay you a dollar for every a hundred dollars of TVL. If you get a hundred dollars of TVL for three months, we'll pay you a dollar, you do it however you want, and the way they do it, typically you say, well, we'll create a token and then try to incentivize people by giving the token. And of course the token is super inflationary and then crashes. And then, it doesn't cover the cost of incremental costs for the people who got in. So I think a lot of people got burned, basically changing those incentives and providing equity. Certainly a lot of people made money, but it was basically spending a ton of money to get a vanity metric without a clear use case, because it was the only point of your defi is that you get to trade defi token. It's very, very circular.

One of the early success of Tezos is you don't see it as TVL, but there are several billions of real world assets and STOs tokenize on a chain. Right now they don't use defi. I think there's some complexity in using those reward assets in defi, mostly regulatory complexity, but I think they are surmountable. They can be overcome. And if that's the case, I think it gets a lot more interesting for the long tail of businesses who tokenize their assets to have access to things like lending, to have access to things like automated liquidity provision. So there's a lot here. And the war is against like, oh, who has a more TVL for trading doge coins. I don't think in the long run is that relevant. There was definitely real innovation on this with Uniswap, but all of the things that we've seen on other chains has basically been, people running EVM clones, and then paying a ton of money for people to copy based contract and then attract liquidity. It doesn't strike me as something that builds a lot of long-term, value for this ecosystem and for their chains.

The quality has been low. I mean, sure, there's a ton of money, pumped by like FTX into all of these Solana DeFi protocol. But at the end of the day, you end up with Soland who will have, a one day votes to, usually liquidated positions instead of actually building a real DeFi protocol. So I remain a little unimpressed with what I've seen here. But the problem with that it's reflexive, like a lot of people agreed in the past years that of course, DeFi is what makes a chain relevant. And so we're gonna care about the chain that have very high TVL. NFTs, I think are gonna have more staying power, especially in the art space. Now, if you look at trading volumes, I would say a lot of the activity and NFT treating volume is still gambling related.

So tech projects, for example, like board apes yacht club, I think it's on the fence. You definitely have some celebrities adopting it, really making a push. You have some sort of, no, it's a club, there's a social status that comes with it. I think they almost pass master in this area. And then you have a zillion clones, which like everyone launched. Some of them launch it with another animal or no, it's not a board ape. It's an okay bear. It's very different. Or it's a degenerate gorilla or some lion. Some of them don't even do that. Some of it is like, no, no, no, what works is apes, let's just do another kind of ape.

So it's an ape that is made of polygon. It's a 3d ape. And none of these, none of these even have a pretense of being a social club or anything like that. Where they are is they're semi fungible. They say, look, they're all very different, but we have three categories. They're very rare, the semi rare and the common ones and they have a floor price. So essentially they trade like a token because they're kind of interchangeable that creates a lot of trading volume. I think some of it is wash trading, but in general, I just saying, it's a game where you trade and you try to predict what other people are going to predict is the one that should have value. So people love this type of gambling games. They've been very popular.

I don't know how much staying power you're gonna have, obviously gambling as an industry has had a lot of staying power, but there's a novelty that came with those and I don't know if the novelty of this very specific type form of gambling is going to last. A lot of the NFTs we've seen on Tezos have been art-related or gaming-related. I think that has probably more staying power because there's a genuine use case here. For arts, it's the best substrate, right? If you're gonna have digital arts, it's easier as an NFT on a public blockchain or an entry in someone's database. And the latter doesn't really make more sense and people do want to collect digital art. People have always collected art. They're gonna keep collecting art and all of art is gonna keep being digital. So for me, that's a no-brainer. Gaming, I think, is a very interesting one as well. The main idea being is that giving studios historically have been known to abuse some of their position. And it's a way for them to basically dye their hands and tell, look, you're in control. You have this, you can actually go to secondary markets and trade those. So that I think is also an exciting application and, you know, Ubisoft launched it's platform Quartz on Tezos for example.

CR:

Nice. I really admire this attitude of building without following the hype. That seems to have been your approach to building Tezos in general. I think, especially in crypto, that's just a very hard road to take.

AB:

I'm starting from the assumption that all this is going to matter in the future.If it's all going to matter, you want to basically skate where the puck is going to be and try to build things that make sense, but maybe that's completely wrong and maybe I should just embrace the hype, get the money, and then, you know, do the crypto spec where you get the money for something silly and then do something serious with it. But the problem with it is that if you do that, you compromise your culture. And that's really hard to get a turn. If you have a culture which basically discourages soft thoughtfulness, at moderation then sure, in hype market like this, you're gonna get tons of money, but I don't know that you can then turn around and put that to meaningful use. That's a very, very hard turnaround to do so. And also, you know, I just play to my strengths. If I wanted to be very cynical and just say like, no, I'm going this is a hype, as much as possible, I don't think I'd be good at it. Set aside the morality of it, I'm not a hype man.

CR:

Yeah. I feel you, I'm kind of the same way with the Defiant. I'm very anti hype and that's sometimes hard with other people who are a lot better at it. But I just wanted to go back to this really interesting point on billions of STOs on Tezos and how that can be turned into defi. How is it being done at the moment? I just didn't know much about that.

AB:

Sure. So there's a continuum in the STO markets, right? On one extreme you have, oh, we're gonna keep doing everything we're doing with a transfer agent and we will hash our capable and every month we'll put a hash, our capable on the blockchain. I don't think anyone does that. That's just like, well, I imagine that one extreme. And then the other extreme you have, we'll have bear shares in the Marshall island, held by a foundation in Panama and some forces that we can have a bear token that actually represents your right, beginning like a hundred percent crypto most products fall somewhere in between. I think you what's important is you need to be able to accommodate for loss. You don't want pure bear assets. Most people do not want pure bear assets for their, u,r holding securities. They want the flexibility of being able to do transfer without having to do a ton of paperwork with a transfer agent. But they also don't want to lose access to their investment if somehow they lose their Ledger device. So you need some form of centralization here. You need some form of KYC, AML, you need some form of market controls, but I think it's possible to do a lot of that and still retain a lot of the flexibility, ad ease of use that you get. And mostly to repairability that you get out of blockchains, but that's a big part for me. A lot of the benefits come from interpretability and today, a lot of these STOs are done on the chain.

So it's nice to save cost for the issuer and for the investor. It also gives you an easier way to do transfers. But the real value I think, is unlocked. Once you realize that, hey, all of these assets follow the same standard. And as a result, if someone wants to build an exchange for these assets, all they have to do is plug it into the standard. And all of a sudden you have access to these exchange. If someone wants to build lending against those assets, they can do that. And they don't have to be the same platform. So this interpretability, this composibility of financial primitive around assets is what's really interesting, but that hasn't built yet. Some people are trying to do this association, for example has been working on using DFI protocols with some tokenized assets.

So there's definitely already experiments in the area, but it will come later. And I also think this is something that works better when the crypto hype dies a little bit. Because the thing is like, if the market is crazy and people are marking money, hands over fist with, with nonsense, why would you bother to build something, you know, meaningful? You build something meaningful, not because you want to sometimes it's because you have to. So whether a lot of the very low lying gains where you just, really is the ends copycat of Uniswap. And somehow it gets millions of dollars, for nothing. When that dies out, think people get their back against wall and say, okay, now I have to do something that's actually useful that people are already really gonna need. And we're not there yet.

CR:

Even after this crash, you still think there's more pain coming?

AB:

I don't think it translates in terms of necessarily prices falling, but I would say in terms of basically,easy money without substance.

CR:

That's true. Maybe a lot of the hypes still need to be washed out. That's a great point. And then the more useful applications can be built. Would be great to see this with STOs on Tezos. Just to wrap up, what institutions are issuing these tokens right now?

AB:

Well, there's different companies we've had of course limited returns on Tezos We've had EquiSafe Global Cap. So there's mobile guns. There's a different set of STO companies that are tokenizing these assets. I think the latest one is an Algae Farm that was tokenized on the blockchain.

CR:

Okay. So do you think the path for, for DeFi on Tezos will be building infrastructure for these tokens?

AB:

Yeah. I think the way to put it is you're doing decentralized finance, but the goal of finance is to finance something. And if all your financing is more finance, there's a problem. Joe puts it by saying, it's like imagine that if you had a NASDAQ and the only thing that traded was stocks for other exchanges, that wouldn't be useful. You need to finance something. So when you say like, look, we're financing an algae farm, okay. There's a use case. People want to get a yield from the farm. The farm wants to get money. They meet-- you've solved the real problem. Aside from that, I think, people want to hold a censorship resistant store of value and they wanna take a loan against a censorship resistant store value they hold. That's a use case, but I guess it's limited as a use case. Surely you can do a little more than that.

CR:

Yeah. That's been the main use case so far, but I agree. There's a long way to go before this is actually replacing traditional finance. Alright, wait, one more question. What makes you defiant?

AB:

Oh, what makes you defiant? I mean, I don't think you can go into the crypto space if you're not defiant in the first place. People say, why do you start Tezos, it's a passion project, but sometimes I joke that because people were wrong on the internet, I'm defiant because I said that proof of stake was gonna work. I'm defiant because I said that governance mattered for blockchains.

CR:

I love it. Awesome. Well, this has been a fascinating conversation, Arthur, really lovely chatting. And yeah, we'll be following Tezo's progress on the Defiant. Look forward to all the big upgrades coming up. So thanks so much again for joining me.

AB:

Thank you, Cami.