Wallet Endgame | Sean Li - Magic Labs | Jane Lippencott - Volt Capital |

[Music] all right all right next up oh first of all thank you Emily live from hype for her presentation amazing visuals and really good breakdown up next have you guys seen Avengers Avengers Infinity War Avengers endgame well do you want better this is the web 3 version wallet endgame how key store Roll-Ups are enabling the future of wallets for humans and AI right now we have sha Lee from Magic labs and Jane lipincott from Volk Capital give a warm welcome to Jane and sha hello everyone thank you guys for being here um I'm going to do a quick poll of the audience before we get started with our panel I would love for you to raise your hand if you already have a crypto wallet Okay cool so a lot of not a lot of beginners in the audience good to know um we have a great panel today we're going to talk about AI user experience and then a couple exciting product launches from Magic um I'm going to kick off by uh a category about AI um AI is reshaping so many parts of our world today um and wallets have not fundamentally shifted to accommodate AI agents or AI powered interactions so I'd love to hear from you are wallets the weak Link in crypto's future especially as it has to do with AI and what is missing from the wallet stack today yeah it's great great question um I would say when we first started magic it was a very simple ecosystem you know where just ethereum um uh one chain kind of one wallet uh very few protocols that people use but now that's evolved quite a bit so now there's huge amount of uh fragmentation um that's caused by the proliferation of chains um but also um there's just so much more apps and protocols so in a way um we solved the wallet creation problem but on boarding there's so much more to go like it's still very difficult for end users to like engage with defi you know which is sort of this use case that lasted multiple Cycles uh like what's slippage what's like yield optimization all of that is very very difficult so but with the um rapid Improvement in AI that could that barrier of Entry could be simplified by quite a bit for just the everyday user like imagine a open AI uh operator that can kind of manage your Finance for you like I could send my friend a th000 usdc and that's automatically generating e and um I think that's a really great way to convey the Magic in like the web 3 experience and to communicate that well Factor security is on top of everyone's mind right now with the recent bybit hack also a recent chain analysis report highlighted that over $2 billion was lost in crypto in 2024 Alone um I'd love to understand from your perspective what the largest vulnerability is today in wallets that might be overlooked by other people building the space or even by your everyday user I I I would say there's two parts uh to this one is um uh currently wallets have very um basic permissioning so so for example uh um there's no uh you're you're literally pushing the assets from one address to another but I think the right model should be that the user is sort of setting uh authorization like Max spend limit here's the duration that you can send here are the conditions and pretty fine rules to send this transaction and that's sort of being solved with a lot of the session Keys permissions um kind of effort um but also I think a lot of um the oversight around Security in web 3 is that everybody is very focused on the kind of the architecture it kind of in a way it's like security theater but ultimately where a lot of these people get hacked is from social engineering um which requires a lot of great like corporate security measures um making sure that no weird things are installed on on your employees computers uh compliances such as stock 2 type 2 or ISO all um just having really great security uh um kind of fundamentals would would help there as well interoperability and composability have both been buzzwords in crypto for the past couple of years um but most wallets are not multi-chain or multi-chain experiences that do exist are relatively weak what do you think true operability will look like in crypto in the near term I think there's there's multiple phases right so you start with you know um a inter interoperable wallet that can go across applications you know that's one of the earliest kind of well factors of crypto um you can use one wallet across many applications so that's solv the second one is chain abstraction making sure the users can interact similarly across chains um there's a lot of great um new technology that that allows that like intense across um even like rhinestone with their Omni account um so there's a lot of great development there that helps with that cross chain interoperability but in the in the future I think um with AI you could actually abstract away the concept of dabs applications completely you're just user trying to get something done and then the on the back end it automates and Aggregates everything for you so so that the user don't have to actually click through all of these um uh dap uis protocols and chains and everything's kind of just automated way so I would say that's like the ultimate version but yeah there like one step at a time right now yeah a lot of people see a future where users uh perform semantic commands and in the back end there's some sort of like permissioned dap store or App Store that agents can interact with directly to get whatever a user wants done done efficiently and in a cost effective way how do we balance this automation Vision that we're all so excited about with control and like user privacy and all like the fundamental guarantees that brought us into crypto in the first place yeah I think I think that's where um it get it gets pretty interesting because you you you if the AI is so smart you don't really want to limit it too much but you also can't use it without the right guard rails um so then it comes in two parts so one there's kind of the permissioning and authorizing uh assets that we talked about um um you know from session Keys um you know like uh but but we we believe that it can get a bit more expressive uh than the currently is like hey you can only make this transaction if the slipage is under this um and I think um that on one hand is very important and second part is being able to actually trust the automation right um uh not just like AI agents but like also telegram Bots or just simple automation scripts um just being able to trust that it's actually doing what it's supposed to do like kind of uh currently with telegram Bots for example a lot of people use it by the way like this Trojan bought processed 16 billion over the past year uh it's like super incredible but users literally have to give it a private key with asset in it so like not very like long longterm but um and also you don't really know that it's actually doing what's what it's claiming to do so I think the verifiable as aspect of automation is um is absolutely key uh as the other components to enable this you mentioned Key Management there how do you feel about seed phrases long term do you think users will be interacting with seed phrases in 2028 I think the seed phrase thing is pretty solved I I feel like you know when when we started in in 2019 we actually launching East Denver um it was the first like embed wallet so you can create a nstone a wallet with the web 2 web 3 login um and since then there's been you know it's becoming really a standard and there's not really a need to kind of you know that that seed phrase problem is solved but now it's more like okay you have these fragmented um embed wallet just sprinkled all across so like how do you develop that relationship with end users how do you um what if the user wants to kind of migrate into another wallet so so I think onboarding is one factor creating this wallet but making sure the user stay within the wallet and continue to use it that's another uh whole exercise aside from the semantic abstraction layer we talked about earlier what are other ways that you think AI agents can explicitly work to improve the user experience uh for people interacting with crypto I I think um I think AI helps really low that barrier of Entry um it's like hey here's some uh like for example like right now discovering uh applications based on what you want to do that's kind of challenging but with AI um they can kind of based on what your input and your intention sort of generate a bunch of like tool calls right and then they can kind of interact with all these protocols um without you actually kind of trying to discover it and then take the action uh on it and I think this sort of is happening in the web 2 world as well where before um you go on Google right it's like okay here's some idea around what I do I try to find a website I try to vet the website I look through the products and everything and then click click click check out so from intent to getting the action done there's like super a lot of areas where users can drop off whereas I think now with AI you can shorten that path uh from A to B by quite a bit so um and I think that applies to web 3 and sort of financial side of this as well definitely agree with you there um I want to move on to Newton if that's okay magic recently announced Newton which is a verifiable automation layer that enables any wallet to create agents on autopilot um there are a lot of potentials for Newton to help improve ux but I love to double click on some of the gaps we explicitly addressed already in this panel how will Newton work to solve those for end users yeah um so so the Newton is based on sort of this um philosophy around a endgame wallet you know what what does wallet look like to um to sort of address this future version where there's like AI right so um I I would say um the wallets should consist of three key pillars uh one is like automatable um for example a wallet could delegate to any other wallet or agent or Bots or automation scripts to take action on its behalf and all of that is done in a way without transferring custody right to the other party so so that's automatable uh we have modular so um uh which means the users can progressively enhance their Security based on their evolving needs um so here we're kind of using the rhinestone 7579 modular smart wallet um that have this modules uh for example you can add like social recovery uh 2fa Pass Key like super easy um and then last uh but not least is uh around interoperable so the wallet is uh can be used across apps cross chains so um with these three pillars um Newton consists of two parts one is the Newton account um that is the modular smart wallet um and then there is the Newton key store which enables the um the automation uh part of this so it's a purpose-built uh rollup um that's based actually on vitalic key store roll up and um and that's going to be able to one synchronize your wallet config and permissions uh to all the chains that support it um and to enable that uh in ensure that any automation by a third party is done um within the user's predefined rules if it's agent it will be kind of like the code of law for agents um and then actually making sure that the script is verified um the automation is doing what it's claiming to do and uh on top of that you can have uh predefined rules and and permissions um that's pretty expressive uh not only it can be stateful or kind of based on like K sentiment or or things like that so that kind of really opens things uh the opens up the possibility and sort of introdu this kind of like permissioning and automation as a primitive right in in onchain so yeah very cool um one of the promises of the ethereum virtual computer is programmability I would love for you to address how Newton enhances programmability for the wallet space yeah I think that comes in um with the automation piece so that is um what we want to enable um so instead of kind of uh user going through all of these like different applications learning the the UI ux and all of that we want to make um sort of Newton a platform where developers or like agent developers even web 2 developers they can easily write like Automation and sort of deploy that with Newton and um you can then uh have the users access um the the programs that's deployed and then the program by default it's like verifiable or verified kind of like uh like little containers um that's deployed uh in the in the te infrastructure um so yeah like we want to kind of shift the Paradigm and do things a little differently because you know AI is eating UI it's solving a lot of these um usability challenges for us so we're kind of of just naturally adapting um to that and and figuring out new ways um that applications in web 3 can be built and how will the user onboarding experience and then experience in the platform feel so so that's sort of where the Newton wallet experience uh that we're launching soon that that will come in is sort of to demonstrate uh this experience for the protocol so very simple still you can log in with like Pass Key email very simple to get a wallet um but then they can be now exposed to like a bunch of like automation things that they can engage with like cop copy trading friends yield optimization um all of that is kind of being handled um in the background um and so so yeah just overall the the vision um is to create this wallet that can self-drive for you you know like self-driving cars you have self-driving wallets that can like trade when you're sleeping and and things like that so so so by now I think we have a good sense of how AI is embedded into the wallet itself can AI agents actually use the wallet and what would their experience look like I would say the AI agent can um in a way like remote control the wallet so so let's say the the agent could use any uh uh wallet themselves um but the user can sort of delegate permission uh to these other wallets that can act on its behalf so this way you're not directly transferring your assets into the agent the agent is sort of remote controlling the the master wallet um with its permissions I think uh one of the things I'm excited for in the next couple of years is for the Paradigm of wallets to shift Beyond just a place where users go to store assets I think that's beginning but we have a lot further to go how do you imagine the shift will look and where will it bring us by 2030 or so I would say um we we kind of got um really inspired with this uh uh this fat wallet thesis by by Deli I think um I think we we might even see a pendulum swing um back into wallets so so what I mean by that is is there's now there's so many chains um doing their own thing a lot of complexity but they're not really kind of building that great user experience um or really connecting with the end users um so so um and and I think that's where it matters a lot that's how you get the old order flow that's how you get liquidity distribution is to form that end user experience and given a um regulatory environment that's uh sort of lightening up um you're going to see a lot of these wallets start to integrate um uh you know all all of this like Defi and being able to um do defy uh and sort of um uh like for example uh infinex would just be able to launch like kaido right away uh when like the centralized exchanges actually will take a while to list them so I think there's like a window of opportunities for wallet to really um start uh gaining traction distribution so um so yeah I feel like in the next 30 Years the users will stay within the experience that they value the most and all the protocol interaction all the complexity that's kind of all being abstracted away in the background so so it sounds like you believe the real opportunity to scale a wallet company is actually ahead of us even though you've been building magic for many years now um if we were looking we were in 2030 now and looking back what would you regret not having done um ahead of this next cycle in terms of building the product uh maybe like targeting a specific user base onboarding the AIS I would just love to hear like maybe if you could outline what your priorities are going forward I think think I think we're doing what the the most of we what we can do right now and the thing that I genuinely want to do the most um and it's a funny story back in uh 20199 or the first version of Magic Wallet was a kind of interoperable universal wallet and one thing that we tried to do is hey like maybe the you know regulatory environment not super friendly but um maybe we kind kind of go a SAS route we build this in bed wallet and maybe apps can kind of customize it and what end up happening is now there's like even more fragmentation um and some of the benefits of web 3 is not being kind of really tapped into from that interoperability so now we're kind of going back to our roots and all the technology is now um ready you know on the smart wallet side on the kind of defi side on the AI side so yeah I think it's like a Perfect Storm to create the next perfect wallet um that sticks to the end uh the endgame wallet principle so yeah that's what um we're working on now well thank you so much for setting the table with us and explaining Magic's new product Newton thank you guys you've been a great audience thank you