How DeFi Builds (And Breaks) Itself I Martin Derka

sessions to go. I want to start off with bringing on Martin from Zurket to tell you all about new ways to think about Defi and security. So, please give Martin a big round of applause.

Thank you for coming. Um, and this is Martin. I'm from circuit and I will be talking about platform specifically about how we build it and how it works itself. So this is define TDL how it was developing over time. This is where we are standing at four weeks ago.

So we had in total roughly 130 billion deposited in D5. But I think that it always is interesting to sort of pause and reflect and sort of wonder, you know, how did it get right? So since we now have 300 million in stable coins and $175 billion of D5 that didn't happen on its own. So how did that happen? So I will go in this talk through some brief history of defi and some significant events that were you know created or maybe also destructive and in the end I will tell you what's happening in circuit and what our response to uh you know is the defi these days is so the best spot to start is actually talking about the very first protocol and for me this is actually the origins of maker they reach into 2015 that's why actually I call it the first protocol spider wasn't uh you know deployed until 2017.

Laker is a collateral debt position protocol. So what it meant especially at the beginning 2017 is that user were able to deposit a single volatile asset that was e and mint stable coin called die against that asset given some volatization ratio. So uh D was the very first widely openly mintable decentral stable coin and maker for us was the protocol we were using as a leverage engine because we had liquidity in ETH we didn't want to sell that ETH but we wanted to use our money so this is what maker uh at the beginning of the existence maker was single plat eventually we developed a little bit more multiplat and so on when I talk about default I I essentially categorize defi into three verticals. One vertical is stable coins. The other vertical is lending and the third vertical is trading.

So let's take a look at for me. The very first D5 protocol focuses on lending is actually it was a peer-to-peer based protocol which operated on request matching. So you say hey I have so much and so much ET and I will lend it to somebody or somebody could submit request hey you know I have uh the need for so much and so much I want for platform will match you and the would be realized when I talk to people about they tell me hey I've never heard of which actually might be true but what they definitely did care about is the successor so in 2020 regretted to a right now is at least for me the actually like so really really big project. Um a is also responsible with this upgrade uh of the system for some innovations. Uh they chose their design through pool based lending which I'm going to talk about in the next slide and they you know introduced something called flash which uh if you have been in DeFi for a few years you have suffered some hacks or if you have been looking at the smartwing field it's not a new term.

So [snorts] uh this is implant. Shortly after implant uh started operating uh a competitor emerged and that competitor was called compound. Compound didn't operate on this uh request patching peer-to-peer type of basis but they had a design based on tools and they still have design based on because these protocols still exist. So LPS would deposit the funds into the pool and then the pool would calculate some kind of interest rate and whoever would come given some collateral all these other tools they would be able to borrow uh assets from the markets. What compound is responsible for and why they are really really important feature is because this was actually the first widely used skillbearing tokens implementation that has existed in DeFi.

So co-ound is really important for us and they are actually important for a few more things. Finally the third vertical for me here is uh trading. So Athereum as a network is responsible for the ERC20 standard right. Uh so if the ICO we created an abundance of tokens there always hands you trade it for that stable coin you trade it for ease but implementing dynamic portable which was actually really really difficult back then for multiple reasons. It was the gas cost and it was also the speed of the network right because blocks were mining within the range of 12 to 15 seconds over time.

So you know this is probably the interval that you would be able to update your your order the book uh at and that's very very slow work. So this was difficult and therefore the researchers at bank were actually came with a new concept called automated market making which was also based on trading. So inside of an automated market maker, parties deposit liquidity into the pool and then anybody can trade against the pool and the tool prices the asset based on some mathematical formula. While Bangor was the very first automated market maker uh actually it's not the most popular automated market maker. The next one that came a little later was sort of took the glory and there are a few more decisions particularly for based on the formula and the way how they price assets and based on some the bottom line is in these early days of E5 and that's why I actually have a TVL chart in there you can see that the adoption was significantly lower than what we are used to uh these things we can actually see that up until 2020 that line remains sort of flat compound didn't develop significant liquidity gains receiving $100 million until 2020 and the whole uh deal itself didn't exceed 5 billions until 2020.

So something needed to change but before something changed, Black Thursday came and this is actually an example of the very first sort of destructive action that I'm going to go through. So what happened on Black Thursday? Black Thursday happened on March 12th, 2020. Both Bitcoin and Ether back then priced in price by 40% within the scope of one day. This is how it looked on Coin Gecko.

Honestly, it's like every Friday this year uh back then we call it back a black. What it resulted in we know from these days very easily, right? It resulted in a series of liquid relations, network congestion and the gas prices spiked up on the network. So you know it was problematic to submit transactions and that again became quite problematic for maker in that because maker had all of these laws right minted. So it hit some faults uh with ETH because price went down under collateralized and those bolts needed to be auctioned off.

And this didn't go easy because despite maker providing some kind of software for running bits for an auction if the auction was secured that software actually only had static gas pricing strategy and because gas prices were rising very very quickly uh bitcoin were mining and there were calls with lot that were being auctioned for $0. Really really beautiful day for some people. So overall, Black Thursday, it was actually the very first example of some systemic problems and it was systemic problems that were caused by leverage, by network congestion and weird execution conditions resulting in lack of participation. Really interesting. So as I said something needed to change for D5 to actually take off and that is what uh sparked sort of uh the D5 summer and what changed is compound inventing liquidity mining.

So somebody at compound realized hey uh we have these people who are depositing liquidity to us and we are rewarding them with something for us but you know we can also create other tokens and you still they will have some value. We can you know put them on the market right liquidity is to develop more that the value will be there and we can additionally distribute these tokens to incentivize protocol uh participation. So this was called liquidity incentives and people were liquid buying it and this mechanic actually became the most prominent incentives for protocol participation all the way up until it was replaced by various points campaigns uh in uh 2023 and 2024. So another very very cool innovation from compound and uh I actually was lazy you know just to just justify my claim that uh this was really really significant I asked GP to list me some protocols that ran uh refining campaigns in 2021 2020 uh and you can see the list on the screen uh you can see some pretty big names and the list is actually short and you know it's much more comprehensive if you ask liquidity mining was a mechanic that created created something else. It actually created a lot of activities.

The problem with liquidity mining was that people wanted to compound their positions. So they were always trying to collect their rewards. What those rewards were additional collaterals and deposits and you know amplify their position, right? But that costed gas and you would always be doing math. Hey, you know, is it actually profitable for me to collect my rewards, liquidate them, really?

So uh some projects came with the idea that uh this can actually be done through smart contract and if we pull all our funds in this effort we will save on gas we will offload the overhead to the smart contract. We will remove the learning curve because it's the smart contract actually to be very very aware of the uh underlying mechanic of the integrated D5 protocols and we will essentially automate the whole thing. So this was super convenient. Very first uh yield farming aggregator was um it was uh shortly followed by farmers and this is when the D5 TVL actually started picking up really really hard. So everybody was farming and these uh yield aggregators were really big because they managed to develop their own ecosystems.

Here I have an example of urine. So in June of 2021 year actually consisted or somehow through series of mergers and acquisitions affiliated themselves with whole bunch of protocols. Sushi as their deex cream as a lending platform. Pickle that was like a farm for me covered insurance protocol. propolis some kind of governance platform taking you know uh stealing all of these and there was also something called iron which was like lat platform which was protocol like the uh kind of intended so aggregators were really big everybody was using them and they are sort of responsible for introducing into our field something that we call normal but maybe it is not entirely normal and it's called loop so here I have an example of a yield loop that to provide it by earn.

You could take your USDC, you could deposit them into Europe, you could uh turn that into yield bearing your USDC and then someone implemented called Apracadabra which was also collateralized that position. So you would take that USDC deposited into Apocracadabra. They meant something called magic internet money which was also a stable coin back to US dollar. So you would want to one swap it for USDC and then you would rinse and repeat as long as the collateralization factors would allow. So this is what deal looping is.

And now this yield looping combined with other mechanics actually led to whole bunch of uh bad things in D5. So this is not entirely moving but the very first or second example of bad are sort of senior controls uh really bad invention. So two protocols to mention one was empty set dollar. The other one was basis fast. They both issued an unqualized stable coin and said hey this will cost $1.

And if it ever costs less than $1, then we will offer to the people to burn these stable coins uh and mint for them some coupons which you know if the stable coin or when the stable coin again reaches $1 or exceeds and price $1, you will be able to burn and you will be able to get yourselves more stable coin and some sort of next bonus. What they failed to explain was uh why this issue had ever happened. It was purely a reflexive mechanic hoping that you know at some point an interesting stable coin will be spark. So it didn't work. Then we had another example of something similar that was called D.

They are actually inventing something we call protocol for liquidity and uh they help program uh for their own because purely inflation program uh and they try to community that uh you know you can hold their you can stay them to get more old or you can sell them if all of us stay them based on some prisoners dilemma is actually best outcome for everybody. It's the 33 that signifi signifies the big scary outcome in here that is going to be great for everyone. But all had no derailers. It was purely inflationally token. It was just being released into the system.

So this mechanic also did work and eventually collapsed because you know once you [snorts] exited the system you materialized your gains and projected losses for everybody else. And an absolute expert in these reflexive mechanics amplified by the looping was the block Luna and their stable coin USD. Again, they had some volatilized uh depth position protocol where you could take Luna volatile asset. You could uh take uh USD uh or mint USD by burning Luna. So Luna's price was going up and the demand for it was going up as well because cost price is going down and there was also something called anchor protocol that was actually offering really really high APYs to USD which really amplified this higher USD spark more demand for USD leading to water and prices going down and that sort of you know kind of worked up until it didn't because all the trust sort of disappeared from the system and USD defect that means that USD get under volatilized uh Luna's prices dropped.

Uh that meant that USD is actually more under volatilized and created this like exploding imploding loop and that actually followed progress and all activity in for a very very long time. Here you can say that this is how long it took for us to develop this liquidity or this TDL and this is the spot where it just you know single dropped. So really just started event later D5 was actually revived uh by the merch which was also the event that developed the staking for their state. So when Ethereum merge happened uh they had a way that immediately after staking your piece you couldn't unstake it. The uh unstaking functionality actually wasn't introduced into bacteria up until one year later with the shell.

So that clearly created ground for tokenizing uh stake ether and those were the protocols of bio with a rocket to them and uh you know later taking in all this huge field of staking and restaking by like buffer rental and actually and this is what helped us revive uh DeFi for where we are today. So where are we today? We are probably on par with the peak of D5 summer or the year 30 of D5. We're still not the state of it and probably you know if you have been following markets we are actually a little bit further from there but we still you know went through a long long long journey. We have many chains all of them have different features.

You have many protocols all of them have different features. Some of them provide E5 landscape themselves. Some of them adds privacy. Some of them add some utility and uh you know provide options. We have whole bunch of SDKs, development tools and we have amazing interal infrastructure.

A year ago I was actually complaining about the existence of whole bunch of chains. This year I'm super super excited. We have a whole bunch of cloud chains that we can build with because that interop is so good. Despite that, we have deployments of protocols that are isolated on multiple chains. Here I'm showing a screenshot of various uh dropdowns selecting what chains uh do what chain do you want to operate.

We also went through some standardization. So we have an ERC 4626 uh for tokenized ws. We have the consensus or agreement uh or you know prevalent opinion that some prevails in D5 actually reach maturity and they might be themselves the primary product market fit for Ethereum. Defy and blockchains internally intersect with our own life. Right?

We have Ether many other companies actually that offer uh credit cards. You can operate with crypto on PayPal and you can actually use it to make purchases on all Shopify stores and as I said before we have $300 billions in stable coins and we have $135 billion in defy applications and deposits. So we really went the long way. Is there any space to innovate still? I say yes because we went so far and we have this you know landscape of things that can be integrated we can actually build better products.

I still strongly believe that it is a responsibility of a defi protocol to protect their users. However you know over time the secondary threats have been evolving together with the field even from bugs over market manipulations and platforms to hacking bridges. These days when people are using blockchain more and more we have the many many fishing impacts and we are responsible for you know a lot of damages and all of that actually aggravates and adds up finally I couldn't ask for more innovations and transparency and managing risk and more responsible defi in general especially in the light of the risk events with street finance what is our response to that and here I'm switching to the national for finance So at circuit we are blockchain at random hacks on uh from the chain uh and we are launching new flagship product. It is a response to the state of defi. It is uh yield product that exists on multiple chains.

It builds on layer zero. So you can deposit funds into it uh you know on any block where the vaults are deployed. system base of liquid and uh primary. Okay. Uh all these vaults they are going to be generating yield.

The yield will be based on lending. Uh the lending is going to be combination of D5 protocols such as a morpho and um C5 opportunities and as I said we have strong focus security. So all will be protected by our system called sequencer level security and data both in the future. This is currently uh in the private interface. We are working with our partners.

We are bringing it into the platform. But if you want to get in uh as soon as uh possible, secure your uh spot on the wait list uh on find.com. And that's everything that I wanted to tell you here. If you have any questions, if you are interested in D5 or building something, please catch me in the back and I will be glad to talk to you.

Thank you.