DeFi Security 101 2025 - The Intersection of L2 and Bridge Security

Transcript

Okay perfect. >> Hey folks, thanks for coming in. So today I'm going to talk about L2s and bridge security and how they intersect. Um it's a cute little topic because it's it's one of those topics which has lots of uh like misconceptions in this space. Um even among OG's and so on. Hello. Okay. Uh perhaps better. Uh okay. So let's start. First uh first of all like I'm going to keep this session relatively small. I I won't talk for an hour. Uh I want this to be more interactive. I'll get some content out

as a like as a starting point and then whatever questions you guys might have any questions uh any ideas we can discuss. So let's start with L2s. So what's an L2? like that's need to define that before we talk anything. So an L2 or layer 2 is basically an extension to a layer 1 chain like Ethereum which in theory should provide the same security guarantees as the L1. Um the reason for L2s to exist is to scale. Uh when we talk about scaling there are at a high level two types of scalabilities like either a vertical scaling scheme where you just have one single system and you make it more efficient scale it uh scale

it as much as possible like Solana or you have horizontal scaling where you decentralize or de uh or you have compute which is distributed and you just keep adding more and more uh distributed compute to have higher throughput. Ethereum chose this way and it introduced the concept of L2s. Ethereum can keep adding more and more L2s with their own capacity and at the end the idea is all of this capacity combined together is going to be much larger than what one single system can vertically uh excelate.

Now for this uh to be really true, you really need the security or guarantees offered by L2s to be the same or almost same as what Ethereum offers. U this is why like we have some guidelines and standards as well. There are uh L2s with stage 0, stage one, stage two that define what level of securities and L2 offer and so on. uh but we are here today to go a bit deeper because a lot of these standards are defined in a very theoretical sense which don't apply to the practical world. Uh for example, all of these standards are meant for the native bridge of an L2. They're meant to

protect uh the assets that came into the L2 through the native bridge. uh but in reality what happens is most of the assets coming in in an L2 right now are not coming from the native bridge they are just coming from somewhere else um so this is an example of base chain for example and the ratios look almost similar the canonical bridge is is the native bridge what I'm talking about 25% of the assets on the chain are coming from the canonical bridge 75% of the assets are coming from somewhere else either external bridges like layer zero um or they are being natively minted on the chain and so on. Now these bridges and assets have completely different

properties. They don't get the L2 security guarantees that the canonical bridged asset uh of a chain gives them. So using them is maybe not the same thing as using that on Ethereum but that's something people miss out on. So when we talk about L2 security, there are really two components we need to talk about. One is the actual security of the transactions happening on L2 itself. Like if I'm on L2, let's say I'm on on arbitum, I send USDC to Alice. U now this transaction has nothing to do with Ethereum. It's just purely happening on arbitrum. um and to guarantee the security of this transaction that this

transaction actually happened with will not roll back and so on that's one sort of security we need from L2s this comes from finality uh finality of L2s optimistic L2s unfortunately is 7 days so you really can't be 100% sure that if if I send money to Alice on arbitrum Alice will still have it like unless you have waited for 7 days for it to finalize but in reality you you would have practical confirmation within seconds because you know like arbitum would give you instant confirmation that okay I've added this transaction to my sequencer and I will

not roll back so as long as we trust arbitum's single sequencer this would be fine the second level of security guarantee you get from arbitum posting the transaction ordering on Ethereum uh which happens I think every 15 minutes. Feel free to correct me but um yeah every 15 minutes the sequencer will post whatever transactions are going to be included um in the chain on Ethereum and then you know like okay now I know for sure this transaction is going to be included. I am not guaranteed the result of it um but at least I know my transaction will be included uh and I can calculate the result myself. I can since I know all of the transactions in

start they are being posted on Ethereum I can run a full node I can know the result what the result is going to be it will not diverge. Um so for practical reasons that those guarantees are are good enough. So yeah when it comes to uh that means when it comes to guaranteeing state transaction on the chain itself like a transfer happening or not the most important thing for the chain to guarantee or an L2 to guarantee is transaction ordering as long as the transaction ordering is confirmed and verified and cannot be rolled back and cannot be changed. The state can be calculated automatically state is fine.

So transaction ordering the most important thing. Fortunately unfortunately right now most of the roll-ups have a single sequencer which means they complete they control the transaction ordering completely. The uh upside of this is there are very few there are very there's very little chance that the transaction ordering would get modified or would get altered or would get reorged because there's these L2s don't have incentives to do it like imagine if arbitum did it once they would lose all of their reputation so they will do everything for it to not happen obviously the downside is that you are trusting arbitum to not do this at least for the first 15 minutes or so

till they post it on Ethereum. Once it's on Ethereum, then it's all good. Then they can't do it. Um, but people don't wait for 15 minutes after sending a transfer. I I don't know if there is any good userfacing tool which even shows if a transfer has been confirmed on Ethereum. It just happens on on the back end. Um, so that's one practical uh aspect of it. Okay, that's for transactions happening on e on on the chain itself and the chain state and chain ordering. Uh the second part is the bridge or asset value. Uh now you're transferring USDC on arbitum but is that USDC worth anything? Even if Alice gets the USDC, is that USDC worth anything? U

that that's the question. Now if that USDC came from the canonical bridge or native bridge of arbitum, it would have the guarantees of an optimistic bridge uh which will have 7-day challenge period and so on. Um so you would have some guarantees defined by people what an L2 should look like uh that the asset will have value. It will have logged component on Ethereum. If you want to withdraw, you can withdraw. You wait 7 days, you withdraw and so on. uh but in reality most of the major assets USDC USDT and so on uh except ETH don't come to the chain through the canonical bridge they come through some other bridge usually circle uses CCI IP uh

USDT0 uses layer zero um so there you have no L2 guarantee like sure you're you doesn't really matter if you're using arbitum or you are using um like a ganache instance in someone's uh someone's laptop, it you'll have the same asset guarantees. Uh if the asset is coming from a third party bridge, if if layer zero is integrated in whatever chain, doesn't matter it's an L2, L3, nothing, L, whatever, it would have the same uh security guarantees. Uh so bridge security is is also important because bridges

are actually like the often misunderstood part but bridges have a higher level say than chain level security. So even if the chain is an L2 chain has its own security guarantees but the bridge provider like layer 0 for example can still rug the chain of all its asset like that are coming through layer zero. uh layer 0 controls the settlement of its bridge not the chain from that I mean uh let's say lay0 is deployed on three different chains ethereum arbitum optimism uh let's say someone deposits 100 tokens on ethereum and gets 50 on arbitum 50 on uh optimism

now it doesn't matter what L2 guarantees safeties optimism and arbitrum wants if any moment layer zero says that the 50 tokens on arbitum are worthless. They are worthless. You can't withdraw them. Layer 0 just wouldn't give you your assets back on Ethereum. No matter what you do on uh arbitum, no matter what force transactions you do, no matter what you do, it's layer 0 controls the finality of its bridge. Um layer 0 does use finality of arbitum to do crosschain. So it basically the chain finality helps the bridge remain solvent and helps the bridge avoid being rugged by other forks but it does not help the

chain uh from being rugged uh by the bridge. If the bridge wants the bridge can still rug the chain. Um there's why like having being having being having being having being having being an L2 is great, having all of those security guarantees is great, but people should also really think about the bridge security part because uh in practice that perhaps is even more important but less talked about. People just say, "Okay, this chain is L2. It's safe. Use it." But 99% of the people that are going to use the chain for are going to interact with components not secured by L2. and they're just independent components. Um there is an intersection in there but two different things.

Some things to note even with canonical bridges uh like optimism arbitum they introduce a concept of custom bridged tokens. Um now custom bridge tokens can actually break L2 guarantees even on the canonical bridge. uh if for example a custom implementation of a token I bridge through the canonical bridge has a function which can burn everyone's balances then there's no way for you to exit like you bring in uh let's say I launch a mod token on ethereum I let you bridge it to optimism using their canonical bridge uh which should give you all the L2 guarantees that you can exit anytime you can use force transactions to exit you can't be censored and so on but I bridge it

through a custom implementation and on the custom implementation on optimism side I add a burn function and I can burn anyone's token and now I go ahead and burn everyone's tokens then magically you can't withdraw you you can't exit even though you came through you came to a true L2 you came through the L2 bridge you did everything right but you can't exit you you can't redeem any of these security guarantees because it was a custom bridge tokens so the these are some edge cases which just exist.

uh force transactions exist um but it's usually limited to just native bridge withdrawals and that to with a delay these don't apply to general purpose transactions like for example if okay you you have deposited your funds on a way and arbitum you want to exit uh the force transaction may not I I think on on arbitum actually they support general purpose but if I remember correctly optimism only supports force transactions for bridge exits which means if if on optimism you deposited money on our way and now optimism is down you want to withdraw your assets you could withdraw if you were holding those assets on an EOA you could do a

force transaction to uh send them send those assets to the native bridge and withdraw from there but if they are in a way you can't do a force transaction to withdraw assets first on a then send them to the bridge and then then migrate so even So you can forced exit. You you will be stuck because there's no you you you're stuck in a way. You are in true L2 but you are stuck in a way. It's it's your funds but they're going to be stuck. Um plus there is a delay in forced transactions usually at least 30 minutes up to I think 48 hours in some systems. And imagine like in volatile environments if if your funds get stuck for 30 minutes, 24 hours, 48 hours, you can't do anything then then you would

get liquidated you you would have bad debt. Anything can happen like in in this volatile volatile market. So all of this combined it it just means L2s have their own even true L2s which confine to all of the requirements defined by Ethereum Foundation and L2 beats and everyone are not the same as Ethereum. There are differences always that people rarely talk about but should u it's not that L2s are insecure they are they are very safe people should use them but at least when we're talking about security especially here people should know what the caveats are uh where the pitfalls are so we can be safe transaction ordering I talked about already it's

it's transaction ordering is the most important thing when it comes to L2 security and people rarely talk about it uh but it it is what defines everything. Like if if you just remove one transaction from the whole list, the result of every other transaction might be completely different. So transaction ordering is really the most important thing. It's not protocol governed. It's really like can be defined by whoever the sequencer is and usually we have a single sequencer. So you should when using an L2 should 100% wait for finality of transaction ordering.

Usually happens within minutes the sequencer post them on on Ethereum and then you can consider okay now my transfer is done. If if you're transferring $10, $100 it's fine. But if you're transferring millions should wait for that confirmation at least. Okay. So we we'll talk about bridges now. We talked about L2s a little bit. So obviously first of type of bridges are the canonical bridges the native bridges I've been saying about um and there there are basically two types one optimistic bridge and one zk bridge optimistic bridge for the optimistic roll-ups zk bridge for the zk roll-ups.

The difference between the two is that the optimistic uh bridges assume everything to be valid uh until someone challenges it. So optimistic bridges have a withdrawal delay of 7 days usually it's a manually defined number can be low can be higher but uh usually 7 days which means if you are withdrawing you would just make a claim that hey I have these funds on L2 I want to withdraw on L1 uh and give me my money the system will then put that claim on hold for 7 days if anyone can prove that your claim is invalid you actually don't have money on L2 then they should challenge they will get some money and that's good if nobody body challenges you can get your funds back in 7 days. Uh this how the optimistic bridges work. ZK bridges uh on the other

hand work on validity proofs which basically say that everything happening must be proven to be correct. Uh you can't make any invalid claims. So with ZK bridges you have uh zk proofs generated for the complete state transaction of the L2. So you can prove that yes you moved your assets from this to withdrawal address and so on and using that proof uh you can withdraw funds on on L1. There's the most basic type of ZK bridge. ZK bridges are much more faster. Uh usually you can withdraw funds within an hour. Um that's basically the time it takes to generate ZK proof. Now honestly like if you put enough money behind it, you throw enough

GPUs, you can have proofs generated within seconds. uh but it's it's always a cost optimization question. It's usually not worth spending that much for accelerating uh proving by 1 hour. So in practice zk bridges takes about 1 hour to exit right now but that's going to come down as we optimize cost for sure. U then we have third party bridges like uh layer zero uh like uh hyperlane u these uh like warmhole. So these this category of bridges are basically sort of like validator based bridges. Uh usually have a component of executioner plus validator where basically uh these

are more assetentric bridges. Uh which basically means that uh to bridge asset from one chain to another. You just do a transaction on one chain. There are off-chain components listening for your transaction. When they see it, they do their work. If they they do add they can do additional check if you want like they might do KYC checks uh they might do they might check your address through against OFAC list whatever you want to define you can define there um and then if they think it's valid they would uh broadcast the same thing on the second chain you want to bridge to and you would get your assets there. [snorts] Um these sort of bridges are usually they can be the fastest ones because they

don't have to wait for finality anywhere. They take on the finality risk uh of a chain if a chain reorg take on that risk. either they would just rug the chain and say hey now chain A you are under collateralized um or they would reimburse chain A whatever but they are taking on the risk uh which means they don't have to wait for 7 days for arbitum to finalize or 1 hour for a ZK chain to finalize they can bridge within seconds within minutes plus they can have additional checks they for a native canonical bridge they have to be permissionless and unopinionated so you can't add additional checks you can't add censorship and so on uh but in third party bridges you can add whatever uh

this why all of the asset issuers usually use them because they want blacklist and whatever features. Um the downside here is that they are theoretically much less secure than a native bridge because they are only for example in DVN uh for example in lay zero they call it DVN hyperain calls it something else wol calls it something else but same concept uh guardian or whatever um that it's it's basically a committee of operators which sign any transaction and whatever they sign happens If the committee says bridge $100 uh to chain B, $100 will be bridged to chain B. If the committee is

compromised, committee can be malicious and drain you. U so there is this risk. Now in practical sense, it's actually very hard to exploit because it's it's not just one person committee. It's it's it's multiple people plus the executioner which usually a different entity controls, layer zero controls uh in their case. So in practical sense they are secure they're fine there might be edge cases some assets which may not have a secure committee but in general uh practical sense they they are fine then we also have uh another set of bridges intent providers uh or intent bridges which are all of the fire right

now uh that I put in third category because they are actually they're not comp competitors to the these underlying bridges uh intend bridges are built on top of canonical bridges and third party bridges like layer zero like intent bridges cannot slash should not have direct mint rights on assets. They they can't uh you don't deposit asset into an intent protocol and uh get new asset minted uh on the other side and so on. No no no intent bridges usually just work by solvers. They maintain liquidity on each side uh and they basically accelerate bridging for you. Uh so what happens is you send let's say you send

them $100 on Ethereum. They give you $100 on arbitum immediately and they charge 1 cent as a fee. The fees are very low right now. Um and the fee they charge because they will uh do a counter trade. they will uh basically bridge funds from Ethereum to other chain using the canonical bridge and actually wait for 1 hour for bridging to happen. So while their assets are are running they will take on the risk of this this opportunity cost of 1 cent and they'll offer you faster bridging. So that's what they are doing. They are a service built on top. Um that's where most of the volume in bridging goes.

Um but yeah they don't really control the assets security as much. Uh okay those are the latest bridges. These are some highlevel concepts I wanted to get out first. We'll take a pause here if anyone has any questions or any topics they want to discuss. >> Okay not. AI >> AI uh AI is not really changing the bridging space but it's making easier for people to understand what's happening. So it's definitely bringing

more u clarity more transparency in the system. Uh because all of this before was just just very very opaque. You you have no idea what was happening. This why like 2 3 years ago bridge hacks were everywhere like you had Ronin hack you had horizon hack everything all multi-chain bridge hack all of the large bridges were being hacked like just two to three years ago now you look at the bridges none of the top bridges have been hacked in last one year or so like lay zero is fine wol is fine after their first hack uh hypian is fine uh uh all of the major bridges are fine because people now understand the risk and are

building accordingly. Okay, these are basic concepts. Uh, okay. Let me then talk about unified bridges a bit. Uh, now we talked about chainto-chain bridges. Uh one thing at Polygon we are doing uh is we are building aglair uh which is a ZK unified bridge uh which tries to combine the best of all worlds.

So first advantage of it is that it's not a one is toone bridge with just ethereum and arbitum for example. uh it is a bridge uh it's sort of a mesh which connects all of the chains that uh all of the chains together that connect into aglar uh which means like let's say if there are 10 chains in a there's ethereum there is uh okx chain there is katana there is zkm and so on um you can move from any of these chain to any other chain directly without hopping through ethereum you you can just directly go from any chain to any other chain uh there's huge step up from native bridges which are just one is to one. Um so unified bridges definitely

would be the future. Now when you introduce unified bridges you have a a hard security problem to solve which is how do you trust other chains. Now traditionally native bridges or canonical bridges have only connected to Ethereum because um all L2 say Ethereum's Ethereum is more secure than me. If Ethereum goes rogue I will be done anyway. So I can trust Ethereum. I can connect to Ethereum without worrying it worrying about anything but I will not connect my bridge to anyone else because I think I'm more secure than others.

What Aglier introduces is a concept of u it's a term called pessimistic proofs very badly named but we are engineers um the concept there is that it basically it every it allows so first let me explain the feature what it allows and then I'll say how how it works it basically evens out the security playfield for every chain and allows any chain to join this bridge cohort this bridge mesh mesh. This means that uh like if if let's say I'm Ethereum, I'm connected to Katana arbitum optimism in this mesh. Um I don't need to worry about any third party random chain

coming in. If I tomorrow I launch mud chain and I connect to mesh, I would not be cont I would not be compromising security of existing chains. the existing change would have still full 100% security and that's because of this uh ZK proven feature pessimistic proofs which essentially says uh and it's a very simple invariant if you think about it and it just says that um the outgoing assets from any chain must be smaller than or equal to the incoming assets in a chain. It's a very simple invariant can be formally proven. Uh but it it's the invariant which is the hardest in practice uh to maintain for a bridged uh bridge network because usually in these

mint and burn bridge networks what happens is you burn on any side you mint on any side. uh if the connection breaks that you are not uh burning enough on one side or if the uh they don't track like let's say uh arbitum became malicious for example they uh inflated my USDC balance to 1 billion whatever and then I burned 1 billion on arbitum and was able to withdraw on on Ethereum 1 billion USDC that would be bad um this is where pessimistic proofs came in come in they pessimistic proofs don't care about the chain state. They don't care if my balance on the chain is 1 billion or $10 or $0 doesn't matter. They say if the chain is saying like from my chain

uh 1 billion tokens are exiting to Ethereum then you must first have a a history or track record of at least a billion dollars uh of tokens coming in. You can't just skip that. Like it's it's basically uh it's a simple accounting book. Every time anything comes in the asset balance of the chain increases, anything goes out balance of the chain decreases. That balance can never go to negative. Um and very very simple concept to manage. Uh you have any question >> is this like UTXO based?

>> Uh repeat is it like UTXO based system? >> No. Um it's not claim based up transaction. It's when the transactions happens. Yes, they can be UTXO but they get collapsed down uh with every transfer. So it it gets summed up into a single number after every transfer. Um so yeah at the end it's just accounting when it's happening. Yes, it sort of is like UTX or not technically but yes uh but as it gets collapsed down it only maintains the final net balance which must always be positive. Uh go ahead.

Hi. Uh you're you're you're right that uh over the last two years most bridge security were outsourced kind of to uh companies like layer zero. But uh have you thought about or would you like to talk about a failure scenario? Let's say layer zero get hacked tomorrow and everybody is screwed. What happens next? and what uh for example polygon would do about it. >> Yeah, I mean that's what keeps us up that that's why like we built aglier to avoid getting hacked and zone but yet talking about lazy because lazy is the biggest third party bridge by far right now. Uh them getting hacked definitely

would be massive. Now it really depends on the level of hack that happens. Um there have been small [snorts] not not hacks but anomalies on layer zero which has been silently patched and uh life continues. For example uh as said layer 0ero takes on the reorg risk if the chain they are bridging from reorggs uh there is a double spend scenario there. It has happened in practice sometimes where okay uh I'm let's say I'm going from arbitum to optimism. I send funds from layer zero uh and I get my funds on optimism and then arbitum reorgs and layer zero doesn't get funds on on arbitum. So layer zero is down uh it didn't happen on arbitum then but it

happened on some chain uh it has happened a few times but in most cases the chains reimburse layer zero so it just happens all under the cover they get paid back they are whole again. uh in some cases when the chain says no you should have waited for my finality I I I'm not giving you any finality guarantees till I post the state on Ethereum uh then layer zero just eats up the cost as cost of doing business sort of so it's part of their risk risk profile these are the simpler or expected losses uh that that are part of the businesses and they happen they keep going on that's fine now a more catastrophic hack uh where like an infinite mug uh mint bug happened or

something like that would be much more dangerous. In lay zero scenario, it would still be uh contained because they have separate DVNs for separate assets on separate chains. So if let's say one DVN was compromised or one DVN route was compromised, that might only let you mint infinite one token on one chain. It won't let you mint infinite tokens, number of tokens on infinite number of chains. Um this means let's say even the worst case scenario their most liquid token USDT0 their DVN gets compromised so you are able to mint infinite number of USDT0 on route um most used chain arbitum for example so this would be the

worst case single hack scenario you can mint unlimited USDT0 on arbitum very massive thing but as soon as it will happen uh all of the exchanges have monitoring for anomaly detection in uh in balances and so on. Uh they would stop deposits from arbitum. Uh all of the third party bridges would stop bridging um assets. Uh layer zero would obviously pause the bridge cut off the arbitum arm as well. Um so it would the incident would get isolated on on arbitum. uh the most damage would happen to the would be to the onchain LP providers that exist on an on arbitum because whoever minted all of this USDT0 will go to all of the dexes, unis swap,

sushi swap and so on and dump their illegally got USDT0 for native assets which can't be frozen like uh ETH uh like D and so on. Um and that would be the maximum impact whatever liquidity is in USDT zero payers on arbitum I don't know right now maybe couple hundred million perhaps um that would be the maximum impact it's still massive impact but lay right now has I think over hundred billion dollars in assets in in in secured so comparatively yes still the maximum impact is very less and that's how it was built Um I'm I'm sure they would be able to

cover the losses like wormhole did when it got uh hacked like 3 years ago. Um so in practice, yeah, if it happens again, I expect Leo Foundation to just reimburse users whatever uh the amount was and life continues. Hopefully people will learn a lesson. However it happened, uh we'll fix it so it doesn't happen again. uh but I don't think it would be as catastrophic as just the chain dying or or something like that. >> Hi uh I have two questions actually. Uh so like one of them is like whenever you when you introduced ag layer the first thing that came to our mind was IBC uh internet of blockchains from the cosmos ecosystem. So I would like uh I would like to know like how do you compare ag

layer in comparison to IBC? >> Yeah for sure. So IBC is just one part of ag layer in some sense. IBC is just a messaging standard. You can do anything. Ag layer obviously has that underlying layer. But the thing that makes it safe pessimistic proof as I was saying is those sort of feature sets are what makes it suitable to be used as a as a universal unified liquidity layer uni universal bridge. Um because then with the security of pessimistic proofs I can connect to as many chains as I want and I'll be safe. If I don't have this guarantee, I would be very resist as a chain. I I will not want to connect to a system where there are less secure than me uh chains connected. Um so that that's why like aglair just you can have

very secure chains. You can you can have very less secure chains all connected together uh having interop between each other, bridging between each other without taking on risk from each other. >> Right? So and like thanks for the answer. Uh the second question that I have is like uh with uh and L2's enhancing the security in terms of moving to next stages as L2 beats has laid out the plan of the road map for L2s. How do you think bridging will change when L2s have like moved on further to different stages and like have developed a lot in terms of their security and sensitive resistance. So bridging already um if we see is moving from canonical bridges to externally

external bridges. Less and less people are uh like so with bridging the thing is you can always have the same asset bridged through multiple bridges. They would they you can bridge USDC through the canonical bridge. You can bridge it through CCIP. You can bridge it through any other bridge whatever. But there is usually one canonical representation which wins. Like if circle says that my real USDC on on arbitum [cough] [clears throat] is the one bridged through CCIP then that's the that's the real one. Um any any other USDC that exists on the chain would sort of be worthless. You

would be able to bridge back and get back your USDC on Ethereum. Sure. But the canonical representation which circle blessed would win which means the asset issuers select which bridge to use not the chain. Um sorry. So as we are seeing more and more asset issuers [snorts] are using external bridges because external bridges give them better support they give them more marketing support. They give them more features and so on. So I think we we we'll keep moving in the same directions where canonical bridges native bridges will keep becoming more and more secure but unfortunately less and less people would be using it using them and people will just move to third party bridges.

>> Got it. Thanks. >> Oh okay. So yeah we started on aglair we have pessimistic proofs. Now all of the chains are connected to each other. You can send money or you can send any assets. um you can send any asset from any chain into any chain anytime and you have security guarantees. The other concept Aglier also brings in is the chain finality um that allows chains to um so Aglar has concept of two sort of two types of chains. One is aglier native

chains which settle to aglar and the other is um we we call them outpost again bad at naming but outpost chains are basically the chains that don't settle to algier they have their own finality mechanism they can settle anywhere the reason we sort of have these two different concepts and we bring them in together is because as said like chain finality which is about transactions happening on the chain itself uh avoiding reorg and so on um is is rate and generally in most cases you would see that the bridges sit above this and the bridges control the asset finality the chain finality whatever um like happens is is irrelevant to bridges

if layer zero wants to rug you despite you being finalized they will rug you now agar combining both of these concept basically means that you have guarantees of both things like it just the bridge is um guaranteed need to be finalizing and the chain is guaranteed to be finalizing to the same state. Uh you can't have one or the other rugging each other. They they're they're finalizing in the same way. Um and yeah, that that's how we solve some of these problems with Aglair. Um and the eventual goal is to again be the be a better native bridge. So I I honestly don't see ag layer as a competitor to layer zero or anyone like

or or hyperlane or these third party bridges. I see it as a better canonical or better native bridge because the level of security we offer are similar or better than what the native bridges offer um much better than what the third party bridges offer and the features that Aglier offers are comparable to what the third party bridges offer. So it it sort of tries to get the best of both worlds. Uh and yeah, that's the goal with a layer. That's a quick summary about the bridge ecosystem right now. Um any other questions we have.

Um not quite sure if it's a fair question. Uh so basically like you were talking about like intent bridging or like third party and canonical. um like I'm not expert in bridging but like I I still wondering like uh for user uh when when we use the intent bridge chaining so basically like we want to get the money in the destination chain and then I release my money in the source chain and is it sort of like safer for user to actually get the money first and then the can canonical bridge >> uh so that's where like the intent provider or solver is taking on some risk >> um [clears throat] that's why they charge a fee >> the intent solver is saying okay like I will have to wait 1 hour on this canonical bridge but I trust arbitum will not reorg so I will give you the

money in advance I will charge you some fees so it's it's just part of their business model plus these intent solvers are so efficient now that most of their bridgings are netted out so they don't even have to take on the risk for example if uh Alice wants to go from arbitum to Ethereum and Bob wants to go from Ethereum to arbitum uh both are moving 100 USDC the solver would just net these two out, give Alice's money to Bob, Bob's money to Alice without taking on any risk while charging both of them some fees. So in practice, most of the bridges now on high volume chains are netted out. The remaining 5 10% whatever you rebalance, yeah, they take on the opportunity risk.

>> That in that sense sounds like for user I should use intent bridging. Okay, for users 100% should always use uh in fact not just Indian bridges, use any bridge aggregator, use the bridge which has the uh fastest bridging time and at least some brand name because you don't want a bridge to just take your money and run away. Uh but yeah, just uh for bridg for users almost I mean the native bridges would still be the most cheapest because they are zero fee. So if you have time using native bridge otherwise use the intent providers. In fact layer zero doesn't even provide a bridging UI as far as I remember it just there no users are going to use uh through solvers

anyway >> cuz cuz like there will be like some other relayer protocol then uh the messaging fail and then I will be like my money will be I don't know where I cannot find it like sometimes it happened to me. So that's why it's like intense sounds like a the way to go for. >> Yeah. Yeah. It's sometimes obviously things can get stuck but those are more technical issues now uh happen much less but if you just message the team they will they will help. Okay, that's great. We have 15 minutes.

Any other topics anyone wants to discuss? Anything about um L2s, bridges, anything? >> Yeah, I have a question. So, in your experience dealing like with institutions boarding on a layer, how do they, you know, manage the risk of using bridges? How do they teams see this bridging risk? >> So, that's a great question. In fact, that that's why like institutions love aglair because they they don't their risk is then just uh isolated to their chain like every other chain in in the system they don't care about um they know like especially with the ag stack that we offer with zk

proofs and so on u they only trust or hook into uh ethereum because that's where aglier settles as well. So yes, we have a trust dependency on Ethereum. If Ethereum rugs, yes, Aglair would would also rug. Uh but they enterprises one, they love interoperability with every chain. They don't know why, but they need to be connected to every chain. They're like, no, our token needs to be interoperable with every possible chain in the world. But they also don't want to take on the risk from other chains. They're like, I want to be able to send money to this chain from China, but I don't want risk there. Um, so that's where Agler comes in. You you can still interrupt with other chains without uh

taking trust or risk. >> So, how does a layer compare to uh other interop solutions like the systems that optimism or uh arbitum are working on? >> Yeah, for sure. So right now what optimism arbitum have are just uh uh single chain bridges just one is to one but they are like uh omni chain bridge from optimism they are making this uh like collective of they're connecting all of these chains together and saying okay you can move from any chain to any chain but the way optimism is doing is there is no security guarantee anywhere every chain is trusting the other chains

uh sequencer to be honest like they are completely trust maximized sort of bridge where if any chain rugs the whole system can be rugged. So that's why they can't let anyone join the join the system. Uh you they will only let trusted sequences join. Uh uni chain would join base they're trying to get base to join. I don't know if it will join whatever. Um optimism will join obviously. Um and they will yeah all of them would trust each other. uh to be honest they could selectively not trust someone like me if I'm base I could say okay I don't trust uni so I won't accept incoming transfers from uni and so on but in general if they are accepting

transfers they are trusting someone um that's that's how it works >> not sure if it's if it's known but uh I listened to a talk about Uh, Hyper British. Hyper British, not sure if you're aware. And how is it compares? It looked a bit similar. Maybe I'm wrong at that. >> Um, hyper bridge. I'm not sure. But in comparison with the optimism bridge or the aglier bridge.

>> Um, well I I don't know hyper bridge. I can say from optimism and uh ag level it differs because on agly I said like even if I'm base I'm connected to agate there is uni chain I can get assets from uni chain with the guarantee that uni uni chain can't rug me because the guarantee is basically provided by aglar even if uni chain uh reorgs and says I'm not sending assets to me anymore aglier would still give me assets because at the end aglier controls the assets um so Okay. Any other questions?

Guess not. I think then we can wrap it up here. Get few minutes back. Thanks everyone. [applause] Thank you very much, Mood. So, you do have a little break and in 10 minutes we're going to have Antoan from the uh Ethereum Foundation.