From Apple & Uber to Creating a Multi-Billion Dollar Wallet Industry

In this EO documentary, Magic Labs co-founders Jaemin Jin and Sean Li trace the company's path from a one-month runway crisis during crypto's 2018 downturn to onboarding over 50 million wallets. PayPal Ventures partner Alan Du explains what actually convinces him to write a check. Sean Li reveals the lesson from Docker's near-collapse that kept him going, and why he believes inaction, not failure, is the real risk for founders.

Jaemin Jin (Magic Labs, co-founder): A big lesson that I learned at Apple while I was there was the obsession with detail. There was a great sense of ownership and craftsmanship. Everyone deeply cared about making it right. Also, when it came to design, building and shipping, there was so much thought put into it.

Sean Li (Magic Labs, co-founder): When you build a product, it's not just to solve a problem. It's almost like storytelling. It's a message. Great storytelling is really about that wow factor. If you're talking about from a product standpoint, you want to identify what the wow factor is and show that to your users as quickly as possible. That's kind of part of the reason why Steve Jobs was very obsessed about the computer saying hello. Whenever you're developing a product, you're kind of almost telling a message to the world about what you're about.

Alan Du (PayPal Ventures, Partner): So yes, the best UI is an abstraction of the UI that doesn't need to prompt you for logins, that doesn't need to prompt you for approvals. It should just happen automatically.

Jaemin: Our ultimate goal for Magic is making technology feel invisible. One thing that I would say has expanded is we're now not just thinking about humans, we're also thinking about AI agents. You're able to trust that an AI agent is making the right decisions and right actions, even while you're asleep.

Sean: Hi everyone. My name is Sean. I'm the co-founder and CEO of Magic Labs. We are on a mission to make crypto more accessible to everyday people. We were known for inventing the first embedded wallet. To date we have onboarded over 50 million wallets and 200,000 developers across various breakout crypto applications like Polymarket, Helium and Immutable, and Wallet Connect to enterprises like Naver and Forbes. But also getting investment from PayPal Ventures, and at the timpe we were able to raise over $80 million in funding throughout this whole period.

Sean: Magic is also the developer behind the Newton protocol, which is the world's first policy protocol to help secure and onboard hundreds of trillions of new assets on chain across AI, institutional crypto and beyond. Our vision is to upgrade the trust infrastructure in the world and help billions of users move on chain.

Sean: I was fortunate enough to have a computer very early on. Over the years I was just really interested in the computer, played a lot of games and eventually was very interested in Starcraft. But at the same time I was very interested in music. I thought that was what I was going to do. I really loved playing jazz music. I played the saxophone.

So I talked to my dad, and as an Asian parent, he found a very clever way to convince me: he said, "Hey Sean, if you do software engineering, you only need to work one job and play music on the side." I was like, oh, that sounds great. I like computers, and jazz is kind of like creating music, art. Why don't I learn building and creating for real in school? So we applied to the Waterloo software engineering program, and luckily I got in. That's how I went on the journey to truly learn to build things.

Sean: At the same time, Waterloo has a very strong startup culture. It was a super great environment where you can learn from each other, you can learn from your seniors. So I was very involved in that culture with the previous co-founders. The project was called Kitematic. We went through a lot actually, so many ups and downs from working on Kitematic. When we launched the Dropbox for your code kind of product, we got like 20,000 developers in one month, and it was growing really, really fast.

We had to raise money. We applied to Y Combinator and we actually got a recommendation there, but after we pitched, we were rejected. So we actually had to ramp down the service to keep it running. Pretty challenging time. We were just kind of new grads at our most desperate moments.

Sean: We got an opportunity from Lightspeed Ventures to come to Silicon Valley for a summer, and they sponsored the whole summer. So we came, and they got all these great mentors to teach us how to build, not only how to build, but to sell as well. Essentially, from that process we were able to iterate on Kitematic. We were able to get the requirement from actual DevOps people at companies who actually just wanted to run Docker for their team locally on the computer.

At the time, Docker was not very smooth. You can't really run it locally on Mac or Windows that easily. So we saw an opportunity. We brought the server software down to a local desktop environment so that you don't have to pay the server cost, and then we open sourced it to get maximum distribution. And so we launched Kitematic. It essentially took off, and we were able to get interest from investors who gave us term sheets, and also companies like Docker who reached out who wanted to acquire us as well.

Sean: I would say the most important part of the journey is that where you end up is never what you originally expect. Progress is not a very linear process. It's actually a lot like art. The more you're super focused on one very specific outcome, you're kind of blocking yourself out of all different kinds of possibilities. But you also can't get swayed too much. You want to commit to some kind of mission. For example, for us it was that we really want to make it easy to build software, something like that.

But with that mission, how you get there is very flexible. It's never 100% clear, and you're also at the same time dealing with a rapidly evolving environment, as well as your own customer needs and your own company's limitations. They say artists work with constraints, right? So there's all these constraints and changes, and I think diamonds come from the friction that you are dealing with in this process. Just over the course of these years, building zero to one again and again, you really get the groove of this.

So when I hit a friction point, I'm actually excited, because it becomes an opportunity for you to find the truth that others wouldn't be able to see. That's definitely a big part of the lesson that I've learned over the years.

Sean: We were able to work very closely with the founder of Docker, learning about decentralized technology at the time by working with him. We were talking about IPFS when I didn't even know about Ethereum. I was like, what the hell is Ethereum? And I was Googling under the table after that. Just something about it really stuck with me. I went to some Berkeley meetups where I met a lot of the OG Web3 folks, and they were talking about all these use cases that blockchain can enable.

Honestly, it kind of felt like a renaissance, where people are in coffee shops and bars just talking about really cool ideas, and these were really big ideas: how it's going to change finance, how it's going to change governance, how it's going to change music, art, like everything. It was super interesting from an infrastructure perspective, but also from a use case perspective. It showed me that blockchain is more than just cryptocurrencies that you can buy and trade and transact, but also these use cases.

So that's when I saw the vision. I got goosebumps everywhere. I remember calling my mom and telling her, "Yo, I'm going to do this. I don't know what I'm going to do, but I have to be in this space."

Jaemin: So Sean and I go way back. We actually both graduated in the same program, software engineering from Waterloo. Sean graduated two years ahead of me. We first both went off in our own directions. He sold his startup Kitematic to Docker, and his product is now known as Docker Desktop. And for me, I joined Uber early in 2015, when the team was still very small, and I got to ride through the hypergrowth phase. So both of us dabbled into Ethereum around 2016 or 2017.

Jaemin: For me, I heard about Ethereum when I was just about graduating university. I heard about this guy named Vitalik, and I heard this word Ethereum, and he was creating this new blockchain called Ethereum. At the time I only knew about Bitcoin. So it really got me interested in looking more into the technology. What I liked about that was, for me, Bitcoin was more about just a value transfer, and at the time it was really fast.

Now it's more considered a digital gold. But with Ethereum, it was adding logic into the blockchain. So imagine there's no middleman, and you can pretty much peer to peer, not just transfer value, but actually add logic to it. Any form of contract, whether that's agreeing on a contract on real estate, or selling assets tied to, let's say, a prediction market, that can all be facilitated through Ethereum. I got interested in that.

And then for Sean, I remember he told me that he was trying a couple of projects on Ethereum, and he noticed that it was really hard to onboard the users.

Jaemin: For me, when I tried out Ethereum I saw the potential of the technology, but I also noticed the broken user experience. Funnily enough, we got to know that we both had seen the same pain point. We actually reconnected in person. That was in 2018. We met at an arcade bar in San Francisco called Coin Op, and it started as casual drinks and then became more like a five hour deep conversation. We kept circling back to one question: why hasn't anyone tackled simplifying onboarding in blockchain?

Why hasn't anyone tried building a very easy login with email? And that's the moment when it clicked. That was a solution to the fundamental problem of blockchain, the broken user experience. So we built the conviction there, and we decided to team up and start Magic Labs.

Jaemin: At Apple, I worked on a couple of things. One was integrating Apple Maps into Siri. The other was directly working on the tokenization pipeline. A big lesson that I learned at Apple while I was there was the obsession with detail, and especially how it was spread through the leaders, like Steve Jobs, Tim Cook and Jony Ive. A couple of things that I noticed: there was a great sense of ownership and craftsmanship. Everyone deeply cared about making it right.

Jaemin: Also, when it came to design, building and shipping, there was so much thought put into it at Apple. Every task or every idea that we had, we approached it with a perfectionist mindset. When I started Magic with Sean, and all the time we were building along the way, one thing that I realized was that perfection takes time. There's no way you can get it right as a small company going from zero to one, to come out with a product that's perfect.

I don't think anyone in the world can do that. When it came to zero to one, we focused on shipping fast, getting the product out so that we could get user feedback, and we focused on really leaning into that feedback, but also trusting our intuition of what the next stage of the product could be. So after multiple iterations, that's when I started to slip in certain parts that we could perfect. And also, with a lot of user feedback, you see these patterns that help you build confidence on investing in certain features to make them very perfect.

So my overall lesson is, it's something that should always stay with you, but it's very important to find the right time and the right opportunity to practice that perfection.

Jaemin: Before Magic came into the scene, the only people who could use blockchain were crypto enthusiasts or technologists. At the time, using blockchain kind of felt like trying to assemble IKEA furniture without any instructions given to you. You had all these unfamiliar parts, and at every step along the way you had to figure it out on the fly. It was so confusing that most people had given up even before finishing the setup. For example, in crypto you have to first download a browser extension, then write down the seed phrase and figure out how to store that safely.

You also have to read and sign these unreadable blockchain messages without really knowing what you're agreeing to. On top of that, there are gas fees and more and more blockchain jargon. It was definitely overwhelming. Where Magic came in, we came up with a very simplified onboarding solution: login with email, phone number and social login. What we did was we took blockchain that was powerful but yet inaccessible, and we made it feel familiar, seamless and secure.

Sean: So the biggest reason why people didn't do that at the time was because the de facto standard was using browser extensions, and there was a lot of dogmatic sentiment in the Web3 space, where as an end user you have to manage your private key 100% by writing it down and burying it like pirate treasure. That was kind of where the space was at. Just from a sentiment perspective, we saw the opportunity: there was a lot of frustration from users and developers having to deal with this dogma, but the vocal minority was really pushing the full decentralization narrative.

What we were working on was extremely controversial at the time, and we actually got a lot of heat on Twitter for even doing this, or even thinking about doing this. So I think that naturally deterred people from working on this.

Sean: Also, at the time, because people were deterred from doing this, they assumed there was no other way to build a non-custodial key management solution, because they hadn't really looked into the solution in the first place. They'd say, "Oh, it's very difficult, people are not going to want this anyway, people want decentralization." Whereas for me it was like, hey, can we do more on the technical part so we can actually get the best of both worlds?

We can offer both a Web2 style login and a Web3 style non-custodialness. We went down the rabbit hole to solve this specific problem, because we also noticed in our sales calls that a lot of the dapp developers were asking, "How is it non-custodial? If it's too centralized, it's too risky, we don't want that." So we actually had to solve this problem, and this showed up in investor conversations as well. All of that pressure actually forced us to solve the problem, because we were already on the path to solve this. We were able to invent the first non-custodial key management patent, called delegated key management, and we ended up filing a patent for it.

Sean: That technology was a key factor in how we were able to complete our seed round, because at the time these huge crypto exchange hacks had really scared investors around investing in any kind of centralized key management solution. But we were able to bring this new innovative tech, which made a huge difference, and that's how we were able to grow very quickly early on, because it strikes the best balance between both worlds, instead of leveraging fully smart contract wallets at the time, which didn't scale super well, or super centralized solutions, which had a lot of friction in terms of adoption. That's how we were able to get to where we were.

Jaemin: So, yeah, in 2018 the crypto market had completely taken a downturn. I remember crypto was being called a scam by the media. Headlines had pulled back, and the ICO bubble had burst. The people who were left over in the space were definitely the true believers of crypto and the true builders of crypto, and we were definitely one of them. Even though the market had completely taken a downturn, we saw an opportunity to solve a real fundamental problem, and that was the broken user experience. Sean and I weren't discouraged at all.

Jaemin: That's something we truly believed needed to be solved. At the same time, at Apple I learned that user experience is actually everything, and someone in the space had to jump in and really start backwards, starting from the user experience and working back toward the blockchain technology. We felt like we were the right people to tackle that. We jumped on it, founded Magic Labs, and pursued our vision of making blockchain accessible.

Jaemin: There were multiple times that I doubted, and Sean doubted as well, if we were taking the right path. There was one time when we had only one month of runway. We did doubt if we were taking the right path, because if we were, we probably wouldn't have ended up with only one month of runway. We were pretty much close to running out of money and closing down our startup. That's the time I doubted it. At the time there weren't actually many people left in the space, so no one told us to quit.

Jaemin: As I said, people had written crypto off. A lot of VCs had pulled out. There weren't many builders left in the space. It was more about just growing with the community of builders that were left over, and we were all true believers of crypto. For myself and Sean, we just felt this great responsibility of needing to deliver the solution, so that at least, even if it was an experiment, it could be tried out on the blockchain and show other developers, and show the world, that there was this seamless onboarding solution that could be tried out.

And luckily, it worked out. But even if it hadn't worked out, I don't think Sean and I would have quit, because we just loved working on everything related to crypto, and we fundamentally believed in the technology.

Jaemin: To keep myself from being distracted by that thought, I continued to talk to our customers and partners, and I knew that they really enjoyed using our product and enjoyed being our customer. That really helped me not be too distracted by the runway problem. At the same time, we felt it was very important to be transparent with our team. At that time we had a team size of probably eight to ten people. They were all working very hard, super motivated, and we really loved working together with them.

We let them know that there could be a chance we may not be able to process payroll for the next month or two, but that our focus was to ensure that we extend our runway.

Sean: I believe in it so much that if Ethereum died, then my company dies too. So I just built as hard as I could, kept going and kept raising money. Fortunately, we were able to close with close to a month of runway left. I had to make sure to take a loan so the money hit the bank without running out. But at the same time, I was still hiring, still trying to find a new office space.

Back then we were in person in SF. I think for me it's like, just keep pushing until the last minute. And I did hear the Docker story too: they were very close to running out, and then the founder created Docker as an open source alternative and launched it at PyCon, and it was a huge success, and really turned the company around. So I kind of believe in those very special moments, and for me, I kept going, both because I believed in the space, but also from what I'd seen and heard, I felt I should just keep going.

Sean: At the time, because the space was really small, there weren't a lot of competitors who had as much adoption as we did. A lot of our barrier was talking to VCs who weren't really bought into the Web3 vision, like most investors at the time. No matter how many developers or dapps we onboarded, it wouldn't convince them. But fortunately there were still a lot of investors who shared that vision and excitement for the space, who really saw the value in us.

When you resonate on a fundamental level in terms of vision and belief, and they like the team, they like the technology, and you're also kind of a leader in the category, then naturally an investment will work out. It was a really tough time for me, where I really had to learn how to fundraise. That seed round was so much different from the pre-seed round earlier on, just in terms of difficulty, especially in a bear market. Actually, we had so many competitors at the time we started, and after the bear market there were only a couple left.

Sean: I think that was the reason. I didn't know we were going to be working with enterprises at first, but I think it was really important to be very secure and follow compliance procedures. So we were actually the first embedded wallet to ever achieve SOC 2 Type 2 compliance, very early on, even though we weren't really working with enterprises at the time, and that standard actually paved the way to us starting to talk to enterprises, because when we mentioned that we have it as a startup, they were impressed that we were able to achieve and hold that standard.

Even today, we are the most compliant embedded wallet solution, with both SOC 2 Type 2 and ISO, which is a payment compliance standard. So that's the baseline, that's kind of a hard requirement to be able to work with enterprises. But also, over the years, we were able to develop a lot of use cases, really well known success stories like Polymarket, Helium, Immutable, and we were able to get Macy's and Mattel, and eventually Naver, but also getting investment from PayPal Ventures. With the PayPal Ventures round, we were able to raise over $80 million in funding throughout this whole period.

Sean: At the time we were working with over 30 million wallets, and now we're at over 50 million, over 200,000 developers now, up from over 100,000. That was a great moment for us to capture the market opportunity at the time, and that gave us a lot of resources to capture the market there.

Alan Du (PayPal Ventures, Partner): I spent probably the most amount of time with Sean, having known him for a long time during our days at Docker. We sat down together, and he walked me through the broader vision of what he was building. It immediately struck a chord with me, in that consumers began to embrace everything that is Web3: real solutions, real financial services tools that can make all of our financial lives much easier. His vision really captured what I think was the ultimate identity problem.

I think Magic's vision at the time was to allow people to complete that journey, or that migration, without having to ask users to change their behaviors or adopt new behaviors. They don't have to go to MetaMask, sign up for an account, memorize the 20 key phrases, and promptly forget them two minutes later. All you have to do is use your original login, and what they do is spin up a crypto wallet for you on the back end, which allows you to connect naturally, natively, to all the Web3 services.

When we talk about what's the easiest adoption or way of adoption, it's when you don't have to do anything differently. That's why I think early on they were able to grow the number of users, or the number of identities, to the millions sort of scale that they were able to achieve at the time, and that spoke volumes about how important user experience design is for the mass adoption of Web3.

Alan: Having conviction is important, and the conviction comes from a few places. One, do you believe that the market is large enough, that they're attacking a big enough problem? And of course, having the numbers that Magic had early on definitely gave us confidence that this is tapping into a large enough problem, that enough people need access to this. And two, is there a technical differentiation or a product differentiation from the competitors?

Sean: And I think the growth that it

Alan: demonstrated, the early customer buy-ins that they were able to get, showed us that there is a good amount of technical differentiation. Like I mentioned, having your own private key management, and getting into some of the more specific features, which I won't bore you with, is important. The nitty-gritty sometimes does win deals, win customers. Just like your UX design really matters: the little things, one extra step in terms of authentication, will probably result in an incremental but significant amount of drop-off in conversion rates.

So I think paying attention to all of those little details, whether it's the feature design, the UX design, all really matters a whole lot. Ultimately it also comes down to the founder, or the founders, themselves: do you really believe they're serious about building what they're building? Like I mentioned earlier, when the market turns, tourists tend to leave, and the true believers, the true builders, will stay, because they're committed to doing this. That sort of grit, that sort of personal conviction that they demonstrated, was also a critical part in terms of getting me conviction.

Ultimately, that's probably the three biggest things: you look at the overall size of the opportunity, whether they have the right to play or the right to win, but ultimately, are they the right people? You can bet on the right horse, but is it the right jockey? So I think those three things really converged on Magic and made me invest in them.

Jaemin: The most important principle when it comes to making decisions between tech and UX is finding the right balance between user experience and technology. I think it really comes down to customer obsession, and this is one thing I learned at Apple: every design, every step of building and shipping, there was so much attention to detail put into it. But more importantly, people really cared about the user experience and their customers. It's more like, when they're feeling the pain, you're feeling more pain yourself. How do you react to that? You go about solving it.

Sean: When you build a product, it's not just to solve a problem. Of course it's got to solve a problem, but it's almost like storytelling. It's a message, and that's kind of like the soul of the product. When I think about storytelling, and how a founder can get great at storytelling, that's kind of why I really gravitated to naming the company Magic, too. It's like a magic show. It's kind of like a performance. If you just showed everybody exactly how the magic works, then it wouldn't be, you know, magical.

Great storytelling is really about that wow factor. If you're talking about it from a product standpoint, you want to identify what the wow factor is, and show that to your users as quickly as possible, and make that a goal.

Sean: Back to the philosophy around technology being invisible: I do envision a future where people can truly focus on what they really want to do, where they're not just dealing with manual things, where it's fully integrated into my day to day life in a way that's really seamless. Maybe there's going to be new hardware that you can kind of talk to.

Alan: Payment, identity, authentication, those things should be ambient. They should happen in the background without us having to intervene. Fully autonomous agents should be able to do things without us having to constantly look over their shoulder. So yes, the best UI is an abstraction of the UI that doesn't need to prompt you for logins, that doesn't need to prompt you for approvals. It should just happen automatically. In some ways, you know, we're in San Francisco, we see Waymos running around, we see Tesla robotaxis being tested.

I don't know if anyone's had the experience of riding one lately. I remember my first Waymo trip a while back. The initial two minutes I was a little nervous: here I am sitting in this car with an empty driver's seat, taking me through the busy streets of San Francisco, which is pretty challenging for me as a driver. Now there's this thing that's taking me around. But quickly I began to just look out the window, looking at the street, and got back on my phone.

I completely forgot I was being driven around by AI. That's sort of the becoming-ambient or abstraction moment that happened for me when it comes to self-driving. The best technology, as fan J would put it, is no UI, it's an abstraction, because you've designed the experience, you've designed the tech, to be so good, so secure, so seamless, that users forgot it exists. So I think, yes, absolutely, that philosophy will continue to guide development, not just within Web3, but also within AI going forward.

Jaemin: We're now living through a great convergence. AI agents are becoming better, and that's how Magic will play a key role in this ecosystem.

Sean: Thank you.

Jaemin: Our ultimate goal for Magic is making technology feel invisible, and my vision still hasn't changed to this day. We continue to believe in our vision being realized. One thing that I would say has expanded is that we're now not just thinking about humans, we're also thinking about AI agents. In five years from now, what we see is that there'll be your AI agent, or multiple AI agents, pretty much managing your money, managing your contracts, booking travel, and doing all the daily activities that you do today, being fully automated through AI agents.

But Newton will ensure that they operate within the guardrails that you have set. You're able to trust that an AI agent is making the right decisions and right actions, even while you're asleep.

Jaemin: Magic already solved onboarding with crypto, and with AI we see crypto being the main rails for agentic AI, and that's where Newton comes in. Newton is a trust layer that acts as a policy engine, defining and enforcing programmable rules such that users and AI agents can safely and autonomously interact with blockchains. Think of Newton as the system on your phone where you get to control which apps can access your location, use your camera, send notifications, and so on. With Newton, you're able to set those permissions and trust that AI will be able to safely interact with blockchains on behalf of you.

Sean: I would say my biggest learning was that progress is not linear. I always like to make this battlefield analogy: you're fighting a campaign, and you don't really know too much about the landscape. There's a lot of fog of war. In Starcraft, you don't really see the map until you explore it. All you see is this hill that you can get on top of. So maybe you get on top of the hill and you can see more, you can remove some fog of war and see more.

It's like, oh, maybe this is not the right hill, I need to go there. And then you go there, and it's a little bit higher, and you can see more visibility into the landscape. But if you don't take any actions to explore this fog of war, then maybe you get crushed, because you don't really know what's going on. So inaction is actually not an option. You have to constantly keep going, keep exploring, keep learning, keep iterating, keep pivoting.

Sometimes it feels like a 180. It's a very chaotic and unpredictable process. Where you end up is never where you start, and most of the greatest companies have gone through a lot of these very drastic pivots. So I think that would be the biggest advice: think of it from that perspective. Commit to your mission, commit to what you believe in, your principles, but be extremely flexible on your strategy and approach. I think that should be the right balance.

Jaemin: If I could go back, I would tell myself to really stress test how deeply I care about the problem I'm solving, and how badly I want to own my own destiny. Because the truth is, when it comes to startups, there will be a lot of pain. When I had joined, I was just excited about building cool stuff, and I did not know I was going to go through so much pain.

But what really kept me here, and I know that I'll stick it through until the end, is how much I care about the vision and the problem we're looking to solve. The other part I would tell myself is, you've got to be really obsessed with the problem, and obsessed with building what matters. If you're not obsessed, it'll be really hard to make it through. It's like a roller coaster, where there will be highs when you feel amazing, but a lot more lows where it'll feel brutal.

So if you have other goals, if you're trying to make money doing a startup, I would advise not to do that. It really has to be something you're so passionate about that you would wake up every day thinking about it. It's not just a job, it has to become your life. So it is a big commitment. It's not for me to scare people, but it's also for me to let people know it's going to be a long journey. But I can guarantee that it's worth it.

Sean: The truth is, a person's personal motivation should evolve over the course of years. I think my excitement about the future of Web3, steady drum beat, pushes me forward, but sometimes, when things get difficult, there are all these personal motivations that come into play. When I first started, I tried to remind myself why I got in: I wanted to develop these skill sets, make more impact, grow to be a leader, learn how to work with other people better.

So in the beginning that motivated me. Later on, we were pioneers for embedded wallets, and I wanted to take it as a new challenge to see if I could create a second album, so to speak, with the Newton protocol. I took that as a personal challenge as well, and we were able to deliver on that too, and to me that's very fulfilling personally. I think as we continue to grow and evolve, we're going to find other motivations as well, but those are just a few examples, and there's so much more along the way.