Chain abstraction is risk abstraction by Radina Talanova | Devcon SEA

[Music] hi everyone uh I'm ADI and today I'll share with you my thoughts on chain obstructions and the associated risks so uh quick refresher what is chain obstruction um and its main goal uh so the idea here is to provide a user experience um that is all about enabling users to use applications without having to interact directly with Bridges or chains so uh what can possibly go wrong here uh this is a very simple example where uh poor implementation of obstruction uh led to a user losing uh some font due to slippage what happened here is that uh user tried to bridge from digent chain back to base and because the UI was using an external Bridge instead of a canonical one uh it first had to slap funds to eat and this is where things went wrong so the user incurred an 87% slippage um and the thing is that this wouldn't have happened if economical Bridge was used or maybe if the user was informed that an external bridge is used or if they had the option to choose uh so this example is trying to show the pitfalls of accountable struct of chain obstruction when implemented on the wrong level so how how can we do it better um there is this example of an intent based protocol and there are lots of different designs but the idea here is that part of the risks are moved away from the user uh so here we have this entity called solver uh that basically once the user deposits funds the solver almost immediately provides the funds to them on the destination chain the idea here is that the user still has to trust the solver to not censor them and the ASC smart contract on the source chain uh but apart from this they don't have to care about uh finality and settlement because those are risks that the so is taking instead of them and uh there is this concept of chain abstracted balances uh this is the third example um the thing here is that if the user only sees uh one total balance uh that they have across all chains they might um not realize uh that for example most of their funds are on a highrisk chain and this is because um the assets that you have uh inherit all the risks of their underlying assets which depend on the bridges and the chains that were used this an example with AP on AP chain but there are lots of different examples and here um the token that is on AP chain uh is basically inheriting security from the canonical Bridge as well as layer zero um so the point that I want to make is that um first of all uh if there are any risks uh that where the US can [Music] um that can still affect the user and they can potentially lose funds they should be shown to the user they should not be hidden and in the example with the chain um abstracted balances uh I believe that at this stage we are not ready to hide the risks uh just because the most of the chains are not secure enough um so chain obstruction should not be risk abstraction in a sense where uh there are risks that um the user is not protected from and they are not aware of them and they might cause funds to uh get stolen or lost but chain obstruction uh should be risk obstruction in the sense that uh we should try to protect the users from those risks and this can be by uh limiting some uh actions on the front end or by introducing some additional uh entities who would bear the risks instead of the users and um as a last thing I want to say that we at L2 beat um want to um I mean we believe in transparency and we also like red warnings so um we introduced these two called Insight where you can check all the risks of your portfolio uh by inputting your address uh so in this example you would see all the underlying tokens of uh a on AP chain and all the risks that come from the canonical Bridge or lezero or the token itself uh and this is mostly targeted towards regular users who want to know how risky the assets they hold are but if you are um developer um and uh you want to make the best decisions on behalf of your users you can also find this useful so uh make sure to check it out once it's life thank you thanks Rina all right we have time for some questions who wants to ask one raise your hand don't be shy you can do it if not I'll ask a question and I know your your questions are better yes amazing hello you and catch um I want to ask like for a simple Finance product um like a very very simple one without all the sophisticated stuff what would you recommend to devs um obviously when you're on boarding completely new be users you you don't want to expose everything what would you recommend devs to yeah so um I think that uh when if the developers want in some way to onboard users without introducing them to the complexity of the chains and the bridges so using some kind of chain obstruction um in case they don't want to surface all the risks because it can be overwhelming they should make sure that they make the best choice for their users so it can be something like limiting the chains they support based on the risks that uh they might introduce or limiting the assets or uh maybe doing some kind of summary of the risk like high uh medium low and then profiling their users uh to show the RIS depending on um yeah their user profile thank you cool we have time for another [Music] question anyone yes here another question is that for the I think the chain abstraction definitely great user experience and also you mention the risk mitigations as well but in general Beyond like the Insight what are there some of the kind of mechanisms which should you guys put in place to increase the user like you know trust the confidence on of the reliability of the abstraction on that uh so the question is what procedures should be followed uh uh primar what other mechanism you put in place to increase the credibility and I'll increase the trust from the user for the account abstraction um all right so uh I think that when implementing chain obstruction and um I guess we can use this example with the uh intent based protocols uh in this case for example if uh the user has to put some Trust on this new entity go over uh we should uh make sure that it's reliable so um maybe we can do something like decentralized solare networks so um for every trust assumption that the user should make we should make sure that this entity uh is to be trusted and it is decentralized enough cool and there's time for one more question let's see if anyone has it I think someone has a question yes yes yes no that's okay uh