{ "episodeId": "SLP764", "speakers": { "stephan": { "name": "Stephan Livera", "role": "host", "tag": "STEPHAN" }, "michael_levin": { "name": "Michael Levin", "role": "guest", "tag": "MICHAEL" }, "olaoluwa_osuntokun": { "name": "Olaoluwa Osuntokun", "role": "guest", "tag": "OLAOLUWA" } }, "segments": [ { "speaker": "stephan", "time": "00:00", "start": 0.0, "text": "Hi everyone, welcome back to Stephan Livera podcast. Today we're gonna be talking about Wavelength, and joining me today on the show, many of you already know, Lalu, aka Rose Beef, he is the CTO and co-founder of Lightning Labs, and also joining us is Michael Levin. Michael is VP of product over at Lightning. Now, yeah, today, yeah, I saw obviously you guys put out this, news about Wavelength, it's your own form of Ark, right? So obviously- You know, long-time listeners will know, but just for new people, Lightning Labs are known for creating LND, which is one of the well-known big implementations of Lightning. You have Lightning Loop and Pool, which, you know, are services that people use for like swapping in and out, and you have, you know, L4O2. So you've got, you know, and Lightning Terminal, I guess. These are tools that we-- that users can do to, you know, manage their channels and manage, Lightning, manage a Lightning node. So, yeah, Yeah, welcome to the show, and let's, let's get into the why, right? So you've already got all this stuff on the Lightning side, what, what's the reason behind doing wavelength, which is an Ark instance, let's say, or a type of Ark?" }, { "speaker": "michael_levin", "time": "01:11", "start": 71.0, "text": "Yeah, absolutely. I mean, first off, thanks for having us. Super excited to be here, long time fan of the show. I think when it comes to why we prioritized wavelength, like you mentioned, you know, we, we have LND, we've been doing stuff in terms of Lightning network Onboard to the Lightning Network, which is exciting, but in a lot of conversations we have, the problem that comes up is sort of the integration cost or the maintenance cost of, managing a Lightning node, right? So you have to set up the node, you have to manage the channels, you have to manage your liquidity, you have to do that actively over time, right? And that is, you know, the, the self-sovereign path, that's a path that anybody can go down, they can spin up a node and, and go down that road, but we wanna The benefits of the Lightning Network when it comes to low-cost global transactions, right, without having to have that integration and maintenance overhead. So we basically wanted to think about a way to put together easy mode for self-custodial Bitcoin and eventually stablecoin payments on Lightning, and that's the goal that we started with, and we worked backwards from there when it came to how do we build out the tech stack from a backend perspective to make this a reality. which is why we landed on things like the combination of, you know, our existing Lightning loop node, which is one of the biggest, most well-capitalized nodes on the network, in combination with this, you know, sort of batching engine, ark-like batching engine on the backend. but Laolu, I'm not sure if you have anything you wanna add to that. Yeah." }, { "speaker": "stephan", "time": "02:46", "start": 166.0, "text": "Well, Laolu, maybe it'd be good if you can add to us, if you can, speak to, you know, the technical choices behind going with, you know, Why was Ark the way?" }, { "speaker": "olaoluwa_osuntokun", "time": "03:00", "start": 180.0, "text": "Like, yeah, like so, I think we're looking at a number of the, the different sort of like approaches. Basically, some of them maybe required like new op code and software, so obviously you can't necessarily bank on that because, like, who knows what's going on? And obviously, like, you know, up until maybe a few weeks ago or like last week, maybe, you know, there was kind of like a lot of uncertainty going on in terms of what's gonna happen with the chain, you know, itself. you know Maybe compared to like making, obviously you can make another chain, but I think that also adds additional complexity because then you have to worry about that other chain, exactly what it's doing, basically, you know, versus just having like a single stack on Bitcoin that can be, you know, fully unified in, in a particular, you know, setting. and, you know, obviously the, the with Ark, the main trade-off, is just sort of like some things around like the way the timeouts work, basically, right? In that, like, there's sort of Active rather than like passively doing something, like, 'cause, you know, for example, like, the way you can view it is sort of like, you know, be-before we actually had like SegWit, people were looking at Lightning Channel, they actually had a Lightning Channel that was sort of like, you know, single directional and would expire basically a-after a period of time, right? And so, you know, for a long time, people were looking at those designs or some very early, you know, versions of that in like Bitcoin J, Bitcoin Seven, things like that as Online at twelve PM, one PM, and three PM, basically the client can, can basically go around that. We also had a number of things on the, client and server side basically around handling this more asynchronous sort of like approach to things basically, right? Because otherwise, like, if you have like a, a TCP connection maybe on your phone, maybe you go on WiFi, you know, to three G or five G, maybe it would break, you know, and that would sort of like cause additional issues around like, synchronicity and, you know, latency and so forth. But And then the client message that says, \"Hey, check your wallet, maybe you have something, you know, there.\" And as Michael mentioned, like we combine sort of like, you know, sort of like what we have is, is the gateway, liquidity network with something like Loop, which, you know, services a large portion of the, network today can, can do, can do, you know, fairly large payments and has like a fairly robust sort of like, inbound liquidity pool from all the other available nodes with sort of like the R component as well, too Basically, then it has the Ark system on top of that, then it also has like the swap layer on top of that as well, too, right? I think one of the unique things that we ended up doing with this too, is that like, rather than introduce like yet another address, like because, you know, people can get confused about that and then those happens, everything is actually just a normal Lightning invoice, right? So we do some clever stuff with like the hop hints that's basically, you know, no matter if I'm sending to, you know, across Lightning, so like maybe And just because, you know, for example, this means you don't al-always necessarily need to route. You could route if you wanted to, basically the client could actually, you know, do the entire thing and maybe, you know, do some modification, but like this lets them sort of like, do something that's a bit simpler and also more lightweight, right, as well, too. Because otherwise, if I'm understanding you there," }, { "speaker": "stephan", "time": "06:12", "start": 372.0, "text": "you're saying like you can stay inside the ark, based on the hop hint, and if, if the other guy's" }, { "speaker": "olaoluwa_osuntokun", "time": "06:23", "start": 383.0, "text": "not in So you would just think everything works like Lightning because it basically uses the same sort of like, you know, because you can see like on the network, layer, like HLc-wise, Lightning is sort of like the SHA two pre-image and some details around like the scripts and so forth, right? And otherwise, like anything that, that can combine those can work effectively and sort of like have like a facade as if it was, you know, just directly Lightning, in, in the system. I think that's very, like, you know, powerful because, you know, for example, That could let us sort of like interoperate and then also have, you know, a shared liquidity across the entire network. We don't always need it, you know, for example, right? Maybe you have like a, you know, five sat payment or something like that. Do I really need to bother all these other nodes on a particular thing or can I go directly to one that's earning similar interest? And I think that one other advantage of that is that like this lets things be a, a bit more async friendly. I think people talk about sort of like offline payments, but that's kind Because you have the ability to sort of like pay someone in the wavelength instance, you know, directly, that means like you don't necessarily care about like maybe holding up an HLC in the broader network for a, a period of time, right? That's something that people were concerned about in the past, like people, you know, maybe wanted to like have longer-lived HLCs be flagged or like they had like an ideas around like, you know, hold fees and things like that itself. But in this case, because I'm not locking any of the liquidity, I can sort of like have that H In terms of what you need to do when you're online is much shorter than, maybe if you were doing like a normal traditional Lightning node, right? Because maybe with Lightning, the sort of like this underlying, kind of like state machine, like the Lightning payment protocol, which is sort of like, you know, sig, sig, revoke, ack, you know, that kind of thing, which maybe is a few round trips. In this case, you come online and can effectively just send, send a signature and then reveal a preimage, right? So now we're able to" }, { "speaker": "stephan", "time": "08:19", "start": 499.0, "text": "shrink that But who are the users you see for this wavelength, wavelength SDK? Is it mainly app builders and A-AI? Is that your target here, or what are you, what are you targeting for here?" }, { "speaker": "michael_levin", "time": "08:40", "start": 520.0, "text": "Yeah, absolutely. I mean, I think like Generally, what we're trying to accomplish here is, if you think about kind of, Bitcoin and stablecoin payments on Lightning as a supply and, and a demand problem, you have the supply side of things, which is developers building applications that, that people want to use, right? And a big bottleneck that we saw when it came to that supply side of things was just the integration and maintenance cost, right? And so by reducing the integration and the maintenance cost significantly and also riding the tailwind of, you know, AI tooling when it comes- To coding and building, we think we can reach a much larger market of potential builders of Bitcoin and stablecoin payments apps, right? So that includes obviously, you know, traditional Lightning and Bitcoin developers, we've already started working and discussing, you know, integrations with some of those, but also, you know, vibe coders who just pull down the information, they point their agent at wavelength, dot Lightning dot engineering, the dots, and just, you know, one shot a prompt and create an application and post about it on Twitter and start to get usage and feedback. Feedback, different things like that, right? so that's another audience that we feel we can target with this that maybe we couldn't before when you had to manage a whole Lightning node. And then finally, agents, right? Like there's this whole, Push, and I think it's early, but we do believe there is something there with regard to, agent to agent payments and agentic payments overall, and we think by like making this, you know, easy, again, you can just point an agent at our site, you can point an agent at the SDK and have them generate their own wallet, then we can really start to tap into the potential use cases there, which is fun and interesting, and again, early, but I think that there has been this transition in consumer mindset around Paper use, which we always thought about and talked about, kind of using satoshis on the Lightning Network for paper use articles or paper use podcasts or, you know, all of that different stuff, which was super cool. But now there's been this massive transition where everybody's used to paper use now, right? Like you are using the Anthropic API and you're doing it on a paper use basis. So we think there's something really interesting about the potential of using, you know, satoshis or stablecoins over the Lightning Network, to pay people globally who are creating AI tools, you know, AI skills, whatever it may be. So that's kind of the audience we're thinking about, it's, it's intentionally broad, and we're excited to, you know, work with all of those different folks to make things for them." }, { "speaker": "olaoluwa_osuntokun", "time": "11:10", "start": 670.0, "text": "Definitely, definitely. Yeah, and like one other thing that I think we're excited about too, and, and one of the reasons why we, like, we opted to basically build our own instance, you know, for example, like as you noted, like there are a few people that have been look, look at some other technologies, one of the reasons we opted to build our instance was also like wanting to make sure we could integrate sort of Taproot assets in like a first-class manner, right? Because, you know, maybe we'd have to like fork some other thing and like dealing with like reb To do the same thing as I mentioned, like I give you an invoice basically, I receive a stable on the other side, or, or, or the opposite, right? I think like the big thing about this, like once, you know, you get, you get working, you try it out, like, okay, you know, you can basically give someone an invoice, you have zero satz yourself, and you can receive, you know, over Lightning, wh-which is, which is amazing, right? Because, you know, for example, otherwise, like on-chain Bitcoin worked like that, This is one of the interesting things, things in terms of like, you know, the way this technology lets you sort of like do that like just in time receive off chain, maybe it's a little bit zero copy, but then you can sort of like refresh that later and then actually have something that's directly anchored onto the chain. And this is great because this also gives you sort of like channel factory like hybrid kind of stuff, which is something we're working on, you know, on, on the, on the side as well too. And as, you know, as, Michael mentioned You, the question you have to like figure out, you know, where to source channel for the agent basically. So I think we have a number of, you know, cool things that been coming out in that, in that area around sort of like the particular pay for use things. I think one, you know, so you said that's very, fairly well suited is actually like, paying for inference for like an LLM API basically, right? Because, you know, for example, you know, today you have some of these APIs, and some of them maybe you require like a direct Like every few days. So I think they understand that kind of like, you know, the difference in terms of like them burning electricity directly for the inference, usage basically. And that we've, that we've seen, like, you know, the demand for this stuff is, is, is astronomical. People making, you know, these massive data centers. And I think like there's really a cool thing where, you know, right now people, you know, t-typically the pricing is maybe like the dollars per million tokens in and out, right? I think we can And also actually works out better for the provider because they're clearing it instantly, they're not sitting on that risk where, okay, well, I, they, you know, I gave them electricity, you know, in return on credit, and we're giving you Satoshi's basically here at this point, and that can also, you know, eventually be, you know, a stablecoin as well, which actually continues to expand the market from there." }, { "speaker": "stephan", "time": "13:54", "start": 834.0, "text": "Yeah. Okay, so let's dive a little bit into the Ark model, right? So, okay, I'm just gonna say my understanding and you, you correct me or you, you update me. So In Ark, there's generally, there's this idea of in round payments and out of round payments, and the idea is that the out of round payment, so I guess let me simplify it this way. In a, in an Ark model, what do I hold? I hold v t xos, right? And so I, I presume it's a similar thing in your model that you hold a virtual transaction output, which theoretically can be exited out on chain unilaterally, but the idea is you stay inside the Ark so that you can say low fees, right? And then the idea is you can make As you said, a little bit more zero-conf level security, and then every now and then the server is gonna run a batch, and at that point you're included in, that's an in-round. Have I got it roughly right there? Or can you maybe explain for me, you know, a bit about how your specific version of the Ark is working?" }, { "speaker": "olaoluwa_osuntokun", "time": "14:54", "start": 894.0, "text": "That's roughly correct, basically, in that, like, you know, the, the server has like an output, right? and like the output, you can say, like, you know, typically maybe like, you know, an output goes back to the client after a period of time, basically, it's actually the output, and the server goes back to the server after a period of time, right? So, the server has like a single output on chain, and like this way you get some of the, the batch savings in that, like, because there's a single So yeah, there are, you can sort of like do a normal batch sort of transaction, you know, and maybe that's, you know, what we call like in routing, or you can basically do some more stuff that's kind of like off-chain, right? So, so today, you know, we do use a similar kind of a concept where you sort of like have like a state chain type approach where you're kind of like splitting this off-chain, and that, that, that, that's nice because you can sort of like do kind of instant receive. But the downside of that is, okay Back in the system itself, right? And this is really cool because, you know, for example, like, you, you can say that a vTXO is already kind of like a channel with the operator, right? It's like a multisig, and you have this timeout, but then maybe we'll spend it, you know, on this other thing. Rather than, you could actually just im- imagine, okay, well, whenever I, I receive for the first time, like maybe it's actually, you know, directly on the thing, I actually get a channel, right? Once I have Basically that let, that lets everyone sort of like route between each other inside of the ark itself. And this is cool because, for example, now when I'm receiving, I don't actually need to like chain, continue to chain these additional instances anymore, right? I can actually just like receive normally as Ellen. And once again, because this is, you know, created using our stack of like BTC Suite, LND, all that other stuff, like we can actually just embed all that very, fairly directly inside the system. We don't actually need a new channel format because maybe that would be cool One of the, one of the, one of the cool things about this is that like it also gives you like a different type of like a refresh model, you know, to in terms of the way this works, right? Because for example, there was something in the past called like timeout channels. It was, it came out like, I can't remember the name, it was some, some pseudonymous person that pushed a mailing list, but I think like, you know, they had a few papers, I think people just didn't understand it, you know, for a while. To be honest Once an output, once like the output expires basically, users would send the coins back to themselves, right? And so what our ended up doing is sort of like, they ended up like swapping for the coins, right? But if you imagine I actually have like a LN channel, and then the, the server gives me another LN channel in, in another instance basically, I can then just send to myself as well too, right? Which is also kind of like a different model, right? Maybe it's like a different trade-off, but at that point, like, you don't necessarily need like the strict round Transaction models basically. And one advantage of this is also that you can actually receive smaller amounts, right? Like one thing that Ellen is good at is receiving small amounts, right? You know, if you look at like a base, a system like an Ark, well, you have like a BTCXO, the question, okay, well, what's the smallest BTCXO, I can have to actually exit on chain feasibly, right? That's a question, okay? Well, maybe it's, it's small if I can actually do a co-op exit, but if I have to go Right. So right now we have a system, on, Wiblend where like, when you receive below one K satoshis, it's sort of like a small, like, voucher system basically, and then once you get, get above that, the server will actually give you a physical BTC. So, and it's actually similar to the way Lightning works, and like, whenever you receive a one, you know, milli-sat or one sat, thing on LN, because the, because like, it, it, it, before it gets to dust, it If I could bridge that to the broader LN network in a way that sort of like is still kind of like, you know, seamless and kind of like gets at some of the old ideas with like trampoline, chain of factoring, and sort of like having hierarchical routing, because, you know, right now like the routing on the, the network is, is fully flat. You basically have like a pub key, everyone has the entire graph. You could imagine something like what we talked about in the password, there's sort of like these like neighborhoods, right? If I can route to the" }, { "speaker": "stephan", "time": "19:12", "start": 1152.0, "text": "neighborhood, I guess obviously the point is this, you know, a lot of this stuff is abstracted away, so I- Yes, yes," }, { "speaker": "olaoluwa_osuntokun", "time": "19:29", "start": 1169.0, "text": "exactly. Right. That's what I was gonna say. Like, yeah," }, { "speaker": "stephan", "time": "19:32", "start": 1172.0, "text": "like, and so I'll give an example, like, I actually just played around early, like, I'm not a giga brain, I'm just a normal, you know, if I could do it, you know, I used, you know, AI Grok and I sent it to Wavelength and I spun up a site and I used a, you know, I set up a It, I sent one thousand, you know, Signet sats, and I was able to receive. And so basically the, the flow for me was like, you know, you create your little, you know, wallet, and you write a little password, and then it wrote, it gave me like the words, and then I was able to just hit, you know, receive, generate invoice for one thousand sats, you know, send them from the Signet faucet, and then just receive them like that, right? So that was kind of, you know, all the complexities The user, let's say, or the, the vibe coder, don't have to think about that. so can you just explain a bit about like what's going on there in terms of like in round or out of round or, you know, in that example where I did a one thousand sat receive into my wavelength, yeah, kind of Wozum browser thing, what's going on there?" }, { "speaker": "olaoluwa_osuntokun", "time": "20:43", "start": 1243.0, "text": "Question, and so, w-w- I think one of the flexibilities of the stack is that like we can sort of like run it sort of like anywhere, right? Like, you know, if you look at sort of like, we have this wavelength SDK, which is like in TypeScript and Python, we can bring the browser, you know, go, I'll also let you generate bindings for the mobile as well, so we have this sort of like, unified code base that can be, you know, have a different frontend facade elsewhere, right? But then if we, if we look Like a short channel ID, which sort of identifies like a channel on chain. There's a protocol that got added into LN like a few years ago, something called SEID Alias. We can have an SEID that's actually not on chain itself, which is sort of kind of like a random, think of like an invoice number basically, right? We, we then put that into the invoice, and then we give that to like another service, right? The service then, you know, does like normal routing, like normal basically, but it gets to sort of like the second to last hop and says The loop nodes, as you know them on, on Taproot mainnet, but then we also have another system where we can sort of like have a fleet of nodes that, like, you know, are individually distinct, but actually operate as a single logical unit, right? Which also, which will help us in terms of reliability and also making sure there's sufficient liquidity. Once it gets to that node, it then says, \"Hey, oh, you know, we have like an, API called HLC interceptor in LND, which I think people, a lot of people use. I VHLC two towards you, right? So it's sort of like bridging the off-chain and like, yeah, I guess the off-chain world basically with normal HLC. And everything is so atomic, it's actually using the exact same payment hash, right? And then inside you as the client, you have a similar logic to what you'd have in LND or any other implementation where it's okay, well, I have an HLC, let me make sure the, the proper value, and we actually also, you know, it's actually also gonna decrypt that onion hop, you know, itself And, you know, the, the, the part that lets you sort of, you know, do this in a seamless manner basically without any additional channels, like, well, the VSLC already exists off-chain and can be extended to you basically. Then you can obtain that and then refresh it, you know, if you need to, or basically just sit on that for a little bit of time 'cause maybe it's like a smaller amount." }, { "speaker": "stephan", "time": "22:59", "start": 1379.0, "text": "Gotcha. And then" }, { "speaker": "olaoluwa_osuntokun", "time": "23:00", "start": 1380.0, "text": "Michael, this would be the--" }, { "speaker": "michael_levin", "time": "23:02", "start": 1382.0, "text": "Well, the beautiful thing about the way it works today, and then all the stuff that Lulu talked about about potentially adding channels into the system, is just like, if you look at the, you know, endpoints, for Wavelength, it's like a very few, very small number of very simple endpoints. It's send, it's receive, it's balances, it's activity, right? And so if, you know, as a vibe coder or something, you wanna integrate or make a wallet, like that's the level Is, you know, how, how do I hit send properly? How do I hit receive properly? How do I see my balances, right? And in the background, we have the, you know, unbelievable talent, the engineering talent at Lightning Labs that's thinking about the best way to do this from an efficiency perspective, from a self-custodial perspective, from a speed perspective, from a convenience perspective, and the refreshes that need to happen and all of that different stuff. And so we have version one of that right now and we feel it works really well, but we're thinking constantly about, One next step is, introducing stablecoins into the system for people who are, you know, really excited about the potential of being able to use stablecoins on Bitcoin and on Lightning. You know, another step, like Olaolu was talking about, is introducing Lightning channels, so that mic-micro payments work more seamlessly so that the, the capital efficiency on our side is, is better. But again, the, you know, those call points, those, those endpoints on the front end, when it comes to that developer interface, don't change, and the expectation is that they Behind it will get better and better and better over" }, { "speaker": "stephan", "time": "24:32", "start": 1472.0, "text": "time. So walk me through, actually, Michael, I think you had a thread recently, might be good if you can maybe just list an exa-- a couple examples of what people have built so far with it, just so people, listeners, can get an idea what kinds of things are people already building on this." }, { "speaker": "michael_levin", "time": "24:46", "start": 1486.0, "text": "Yeah, absolutely. We, we saw somebody who built, a fully self-custodial Wavelength, wallet within a, Telegram, and which was super cool to see. on that side of things, we saw a couple people who just built like Wazir style wallets, on the web. I actually took some time and vibe coded, an integration with Buzz because we launched on, on the same day as, Block released Buzz, and so I was just playing around with, you know, giving each agent in For instance, through Goose and ACP, a wavelength wallet, and having them be able to pay each other and ask for tasks from each other with regard to having like a sort of, overarching master agent, that was sort of farming tasks out to sub-agents. we also saw somebody on our community call who used L402 and wavelengths to basically, allow for clients to get access to particular like financial data that was behind an aperture endpoint and required L four o two, so you could ask for, you know, information about a ten K or something along those lines and pay that, L four o two invoice using Wavelength. So we've seen a lot of people, you know, starting to experiment with it, and we wanna hear from all those people. We have community calls, so you know, look out for, for our announcements on Twitter, reach out directly to us, obviously as well, 'cause the feedback that we've gotten thus far has been super helpful. We've already had a, a And just to improve the system based on, you know, the feedback that people are giving to us. Yeah." }, { "speaker": "stephan", "time": "26:21", "start": 1581.0, "text": "and so one other question I've got, like, how, how-- So we were talking a little bit about like offline payments, right? Or asynchronous payments in that model. So let's just say, as an example, you're doing like a web-- a Wasm thing, and you try to receive, but then you like-- I guess, I guess as an example, how, how do the, you know, if you, if the user closes his browser at that point, but ultimately the underlying, you know, web server is still running, is that still able to take A, a wavelength payment, even if, you know, his browser is closed, because ultimately it's about that server being up, right?" }, { "speaker": "olaoluwa_osuntokun", "time": "27:01", "start": 1621.0, "text": "Yeah. Good question. So like, you know, if we take of the case where it's sort of like wavelength to wavelength, like, you know, we're both on the system basically, it's, it's, you know, fairly straightforward because, you know, for example, in that case, like, the payment is just sort of like pending until you come online to actually then, you know, pull it and withdraw, right? And that's I think a lot of like efforts, I think we can sort of like see wavelength sort of like these applying platform infrastructure to let people build these future walls on top of, you know, which I think is definitely a really cool thing. On, on the LN, domain, it's kind of like similar to sort of like how things work now, maybe the payment would just be, you know, there for a period of time until the, the user then came back online basically, right? On, on, so for, for example, like if you're doing like traditional mobile phone It's not sort of like, you know, needing to you to be alive for a period of time, because typically on these phones you can't really be online for that period, for that long of time. So then in that case, you can then actually just, you know, wake up the phone for that period of time and then settle it. So I think the mobile experience would be a little bit better, but yeah, if like the computer is completely offline, just like in, you know, other, you know, cases, and this sort of like, this is the case for To like do the final settle, but we're trying to like make sure that's like shrunk as, as much as possible, in the future. Gotcha." }, { "speaker": "stephan", "time": "28:26", "start": 1706.0, "text": "And then, so I guess from a builder perspective, it's sort of like just understanding which piece needs to be online in order to be able to still receive the payment. And I guess it just depends on what people are building, because maybe they're building a website and they just wanna be able to quickly take Bitcoin payment for their own service, versus they are building a website where each user of the website wants to have their own underlying wallet Complexity in managing that, right?" }, { "speaker": "olaoluwa_osuntokun", "time": "28:55", "start": 1735.0, "text": "Yeah, yeah," }, { "speaker": "stephan", "time": "28:56", "start": 1736.0, "text": "exactly," }, { "speaker": "olaoluwa_osuntokun", "time": "28:56", "start": 1736.0, "text": "exactly. Yeah, and, and so, and, and as soon as I mentioned around sort of like the things around wake up and so forth, that's, that's all things we're gonna continue to build into basically our SDKs on like the TypeScript one and also the mobile one as well too, so people can basically just reuse all that functionality and make sure that we sort of like have like a rock-solid base they can then build on in the future basically." }, { "speaker": "stephan", "time": "29:11", "start": 1751.0, "text": "Yeah. And so then, all that stuff is open" }, { "speaker": "olaoluwa_osuntokun", "time": "29:12", "start": 1752.0, "text": "source there too, so the people can also extend it, you know, and-" }, { "speaker": "stephan", "time": "29:17", "start": 1757.0, "text": "I see. Yeah. and then I guess, talk to us a little bit, 'cause we've spoken a bit about the, like, the Wasm kind of web browser style. What about on the other side, like, if you, if you're going for like a mobile app and you wanna do Wavelength inside of that, what does that look like? Can you talk us through a little bit of that?" }, { "speaker": "olaoluwa_osuntokun", "time": "29:33", "start": 1773.0, "text": "Yeah, so like, so we have like the wavelength SDK, which sort of like the TypeScript one, and we also have wavelength, dash mobile, which then also like, you know, wraps our GoSwift and gives you sort of like native, you know, Swift and also Kotlin, kind of like the more, you know, modern languages for iOS and Android development, basically. And you can basically drop that into your application and then start using that. And I think just even like early this week, someone on our team started to develop their own iOS mobile app, basically, found some bugs, LRPC service area, like it's quite large, like there's maybe forty, right, across different subservers and things like that, but, you know, at that point, like you're doing like node management, you're doing fees and things like that. In this case, but in this case, it's just more about like sending and receiving basically, so things can be much more simplified in that case. So I think people will worry more about like maybe things on the user experience side of things, maybe on the UI basically, and then other things around robustness to make sure that things work effectively. Other, other stuff, you're basically just worried about the presentation to the end user, and then if we're doing our job, everything should just be seamless on, on, on your side. So let's talk a little bit in particular on the mobile side." }, { "speaker": "michael_levin", "time": "30:42", "start": 1842.0, "text": "Oh, sorry, I was just gonna say, in particular on the mobile side, like we feel like we do have the, you know, releases that Laloo mentioned, we wanna collaborate with people who are building mobile wallets because that last mile, with regard to the wake up when it comes to push notifications, and hopefully being able Be done in the, in the background. so we're super excited to like collaborate with mobile devs and, and sort of make sure that we're building out that last mile to make that as seamless as possible." }, { "speaker": "stephan", "time": "31:13", "start": 1873.0, "text": "Yeah. what are some of the trust assumptions on the server? Like, can the server steal your funds or can the server like just deny you but not take your funds because it's self-custodial?" }, { "speaker": "olaoluwa_osuntokun", "time": "31:26", "start": 1886.0, "text": "good question. Yeah, I mean, so like the system has unilateral exit in that, like, you know, if you have, like, you know, off-chain state, basically you can get back on chain, right? You know, maybe it'll take, you know, a few blocks, maybe you have to pay some fees, but that's sort of like a cornerstone of how this works in general. And like, you know, some of the other systems out there may be like have a thing where, like, you're not actually fully settled in any given time, but like, you know And kind of like have some sort of like, you know, qualification thing where they're shooting conversion for it out there itself. the one thing is sort of like, whenever you're receiving, doing the off-chain initially, it's sort of like, it's kind of like a zero-conf thing, similar to how people had like zero-conf JIT channels back in the day, basically where like you would get a JIT channel and then, then it's, it's there. The difference for this one is that like rather than needing to, you know, allocate, you know, They went offline, now I have that dead weight, you know, lost capital basically that I can't, I can't really do elsewhere. With this system, you can actually kind of like move it around in a more of like a, you know, a pool fashion basically. So, and as you mentioned, like some of the stuff that we're doing on the LND side can let us sort of like eliminate more of the, zero-conf nature of that, like in that, like, so maybe you could potentially double-spend that, right? But if you just have like a channel So are working on some different, you know, variants of the system, ones that like, you know, maybe use a different type of system for like the atomic refresh, maybe one that like has like a different type of way to, you know, manage this off-chain state basically. But yeah, so that, that, that's the one trade-off that like, you basically have like a zero-conf kind of, you know, confirmation or security when you do the initial receiver, then you can upgrade that, but then, you know, things around the channels can actually let you, Matter if" }, { "speaker": "stephan", "time": "33:25", "start": 2005.0, "text": "you need to. Gotcha. Let's talk about fees. What kind of fees can users expect to pay for this to use Wavelength?" }, { "speaker": "michael_levin", "time": "33:36", "start": 2016.0, "text": "Yeah, so initially, like with the alpha, as it currently exists, it's basically a one bips service fee on Lightning transactions, plus just normal routing fees. So we're not actually charging, Lightning swap fees, and then there's kind of standard on-chain fees when, you're touching the chain. This is for alpha, like we really want people to be using this and getting feedback on this. you know, as we evolve and the product matures, like I think we're gonna be making some determinations about what these fees should actually- Actually, be over time, you know, based on the, the type of usage that we see, right? Are we seeing more usage of people sending out to the Lightning Network? Are we seeing more usage, kind of intra, within the wavelength instance, which there's no fees on right now, like if you're sending to somebody else within the wavelength, then there's no fee on that, right? So I think we really want people during this alpha phase to, you know, build things out, help us understand the way flows are gonna work, work with Determination with the, you know, partners that are building with us about the, the best way to create those." }, { "speaker": "olaoluwa_osuntokun", "time": "34:41", "start": 2081.0, "text": "Yeah, one thing that we dove into, like, there's sort of like an economic theory behind, like, kind of like Ark fees in general, right? Because, for example, like, like the, the server, like, this isn't like an exact term, but like, they're kind of giving you like sort of like a short-term loan in order to receive, right? You can say, right? Because, like, for example, every single round, like, Right? So like, you know, if you're kind of like going for a mode to avoid sort of like people just rewriting on the system basically and like absorbing all the resources, you need to have some pricing of that, right? So, there's sort of like a curve where like the sooner that you, sort of sorry, like the, the later that you refresh something, basically like the, the lower the fees are, right? Because for example, like let's say like you have something and like it expires in seven days, right? If you refresh it on like, But, but let's say you refresh like the, the record the day before or maybe the hour before you actually get the export, well, the server now at that point can get their capital back very, very quickly, and this means they have to like have less float in the system in order to accommodate all the users, right? So we have a system like that where it's sort of like if you refresh sort of like at a later, later point, it could be sort of like very, very low or even like free potentially, basically, right? And the other thing too is that like whenever you have Two output or two and two out, basically depending on what's actually going on there, right? So that, that, that's one component that's sort of like unavoidable and like it's one of those things where like people, you know, advertises like, are, that maybe, maybe remove liquidity concerns, like, not necessarily, it sort of like concentrates them, but lets you sort of like amortize it over time and, and develop sort of like addition, additional kind of like, you know, models to actually, you know, have this function effectively. You know, you could have it Is sort of like, if you, if you move it very, very late in the cycle, I think some of the other people would do this as well, it can end up being free, but otherwise, there's kind of like a cost curve that we're looking at basically. Different, you know, it depends on like your assumptions around like the, the payment size, number of users, like what they're using it for, are they doing, you know, kind of like a passive hold, are they just sending a lot basically? So it's one of those components I think we'll, you know Yeah." }, { "speaker": "stephan", "time": "37:05", "start": 2225.0, "text": "Yeah. So as you're saying, it's like, it, it's kind of, it's a different story in terms of the cost. It's maybe a little bit closer to a leasing model, and I guess the larger your balance and the longer you wanna lock it up for, the high, h-higher your cost, and let's say the less your balance and the shorter, like if you're doing the refresh very close to the edge or the end, it's a, it's much cheaper. So I guess that's kind of loosely, so maybe loosely speaking, it's kind of It might make more sense for spending wallets, like obviously hodling larger amounts, bigger, you know, bigger amounts, long time of holding, you should be putting that in your, you know, in your multisig or whatever, that's a different, in your cold storage, right? Whereas this is more about spending, and then that also has implications on if you are a builder, what kind of products are you gonna build with this? It makes more sense for, let's say, the spending use cases, let's say, that's, I guess that's how I'm understanding it There's all these like AI attacks going on right now. So Lalu, obviously we've got to get your perspective on all this, right? Yeah. We've seen, probably the other big one, you know, in this world is Boltz dot exchange, right? Recently they went down and now it looks like they may be handing over to somebody else. Wanted to get your input, your thoughts on this. Are, Bitcoin, and L2 and open source, you know, services and providers under attack? AI assisted attacks? What's your, you know, what" }, { "speaker": "olaoluwa_osuntokun", "time": "38:26", "start": 2306.0, "text": "do you Like, I feel like I've seen this come up for some time, just once, like, you know, I just started to grasp like the capabilities of these models. Basically, you know, even aside from maybe like the hype or like hacking techniques or whatever else, like, you know, these things have so much particular knowledge. I think in the past, like, in order to like do auditing or maybe some security work, you had to have like a very, very careful mind, like, going through every single line of code, drawing out different scenarios, basically, like, typically, like, you know Right? I think, you know, up until now, I think people hadn't really think, think about like, make sure they have the expert defender, or maybe the models they were using actually wouldn't allow them to actually, you know, do this effectively, right? And this is something I think we've been sort of like, you know, thinking of, you know, for some time. We've also been doing our own background scans and there's different ways you can sort of like, you know, have the model look at a little bit differently to maybe actually have some Because like they didn't have like, the team or the, the proper way to actually, you know, classify these instances. And I think one thing that, we've seen the trajectory of is that now, like, now that you have these open models that actually have, you know, minimal guardrails basically, which actually let you defend effectively, I think that's caused people to basically wake up and, and make, ensure that things are, are functioning properly. You know, it sounds a little ridiculous to say, but like, you know, there are a finite amount Find vulnerabilities basically. I think once, once that, once, once that's done, every single new release we, we basically have additional height in kind of like requirement around exactly what's going on, additional auditing, additional reviews. The one thing that I think like, you know, definitely has shifted is sort of like vulnerability disclosure timelines, right? And sort of like how people are around with this stuff, right? Because, you know, I think before, like, you know, I think a lot of products sort of relied on the fact that you had to be sort of like a specialist to actually Right? Someone could probably try to reverse engineer stuff a bit more easily, right? So I think there's kind of like a, a trade-off between, you know, information that's out there basically versus trying to like, you know, protect your users some, some degree, right? So I think maybe it'll be a bit just messy in the short term once people just realize, okay, like, hey, we need to, we need to get this stuff out there, and people definitely should, you know, should be patching, there's a new version because, for example, we have Actually, even if you were a developer yourself, you kind of have to rely on the maintainers to some degree to make sure everything is fine, right? Like, you know, but now, like, you can actually verify yourself. You can trust and verify because, like, the first thing now people do is, okay, I have a new software, let me ask Grok, let me ask Claude, let me ask, you know, K three, like, is this secure basically, right? Whereas before you have to pay an auditor ten thousand dollars or whatever else. So I think, I think this Over time, I think defenders will have better and better tools, and everyone's gonna have basically passive scanning capabilities, that they're running all the time on, on, on new software releases basically." }, { "speaker": "stephan", "time": "41:35", "start": 2495.0, "text": "Yeah. So I guess you're saying, yeah, as you said, there's gonna be this disruptive period. Hopefully, most of the low-hanging fruit gets, at least from the attacker's perspective, gets cleared out, and we, we get those bugs patched. And I guess maybe we're gonna be running on faster update cycles for a little bit now, because everyone's just gonna be quickly patching Curious, do you see, like, if you look at the kind of the Bitcoin industry and the open source and all these things, do you think as an industry we weren't focused on security enough, or is it just more like people didn't really know what was coming with AI and it was just difficult to predict, like, how this would come?" }, { "speaker": "olaoluwa_osuntokun", "time": "42:16", "start": 2536.0, "text": "We're always cognizant, like there was always like the air of, like, you're making Bitcoin, Bitcoin software, so therefore it's not like normal software, right? And there's no auto updates, like, you know, you're basically be, be, be, be, being a bit more careful. I think one of the things is sort of like, when software gets more complex, it's harder to, like, you know, in your mind, really go through and, and handle all the different control paths and so forth, right? And so for example, like, I think what Whenever you're working with agents, like they'll wanna just add stuff, right? A bunch of different checks, things like that. Like you have to separate, you basically say, \"No, let me like, you know, reduce that basically how that to a more, or more minimal sense, basically. \" So I, I think people were paying attention before, like, you know, it's one of those things where you can tell, \"Okay, well, how did people miss the Coldcard thing? \" Right? I think even people like Ledger had looked at, you know, their, you know, def-def-def-def-define the way, the way they, they thought it was. but yeah, I, I, I think things will just continue to get better basically. it's one of those things where, like, yeah, like, security has always been hard basically, and you always had to have that specialized instances of individuals. And, another thing in the past was sort of like, at which point, like, did you say, like, we're secure enough basically, right? Because you, you, you wouldn't, maybe you Like a defense, you, it's, it's a mat- it's a matter of like how much you're willing to spend on tokens for this, basically, right? Because, you know, if you're not spending tokens on, as a defender, the attacker's spending the tokens, basically, right? I think this is all-- This will also end up, I think, shifting bug bounty programs. I think in the past, you were basically paying individuals for like their work for reporting some, some vulnerability in your software, basically, but now it's one of those things where it's like, okay Maybe it was a bit more indirect around like number of engineers, number of hours, whatever else, now it's literally sort of like the amount of tokens that you have, and that maybe this even less organization sort of like budget this stuff a bit more effectively, right? Because we're gonna spend ten thousand dollars on releasing this thing, you know, upfront to maybe you never have to have a bug bounty in the future or something," }, { "speaker": "stephan", "time": "44:31", "start": 2671.0, "text": "right? And doing more scans, yeah." }, { "speaker": "olaoluwa_osuntokun", "time": "44:32", "start": 2672.0, "text": "Alright," }, { "speaker": "stephan", "time": "44:33", "start": 2673.0, "text": "so bring it back to wavelength, let's, I guess, where do you see it going? Give people something to, you know, be optimistic about. Where, where do you see some opportunities here? you know, whether it's, you know, using Wavelength to, you know, set up to take, you know, do a business or try and set up, you know, for Lightning payments. where do you guys see the opportunities?" }, { "speaker": "michael_levin", "time": "45:00", "start": 2700.0, "text": "Yeah, absolutely. I mean, I think for me, like, I've worked on kind of platform-style products my whole career, right? Which is, is, is sort of a-- it's a bet on if you make something easy for builders, like, it's a bit of a bet on human ingenuity, right? And so I think by making the, the product as easy as possible, literally anyone can build on it. Anyone can start with a POC and just kind of see what the reaction is and, and sort of, sprint in the direction of Of feedback, right? Especially with the, the AI tools that we have today. And so I think the opportunity is kind of a bit of a bet on, you know, human ingenuity and people who are thinking about Bitcoin and stablecoin payments, people who are thinking about a global payment network that's, you know, low cost, fast, you know, all of those different things. There's so many different things that can be built on top of that, and I just want to make it clear that, you know, anyone can start to push in that direction. The areas Of things where you have an agent that has a wallet and they're going out and autonomously paying for things, and then paper use in this AI world in terms of people just being used to the idea of, you know, paying per use on an API, potentially we have things like paying for tools or paying for skills down the line, so yeah, I think those are, you know, interesting areas for opportunity." }, { "speaker": "olaoluwa_osuntokun", "time": "46:23", "start": 2783.0, "text": "Yeah. Two is like once you, once you have separate assets on them, you have like stables on there, that also gives you like a new type of wallet basically that you can kind of like seamlessly move, you know, in and out. And I think one thing about what we're trying to do, we're trying to make sure it's all like on the Bitcoin stack, 'cause then, you know, obviously you can do other chains, you know, for example, I think the other day Solana like halted, you know, yet again. Basically, we want to make sure that we have that type Ties, they have like a route into a system basically, kind of like similar to kind of like a trampoline style thing, because now at that point you can sort of like have like a hierarchical route. You imagine like, imagine like every invoice has like, you know, two pubkeys, kind of how it is right now, like the, the, the sort of like external pubkey and the internal pubkey basically. You can route to the external one basically, and then, then from there they can do the internal one within that too, because for example, if this thing is taking on more and That can then route into the end of other individuals, right? And if you extend it further, you can even allow sort of like routing between the neighbors directly as well too. So I think, I think it all, all can, all services can maybe reinvent some of the research that was going on in terms of like counterfactories and trampolines and things like that as well. Now we have a more direct application. I think the other thing too is that like, it's sort of like a platform like Michael was saying that like, now you can sort of like do these like, you know Something like this, and actually be, be much more chain efficient, 'cause now you actually just have it on chain, and then you can also have, maybe people selling, you know, stablecoin on this as well, too. So I'm very interested in sort of the combination of, of this stuff, and also now because wallets are easier to make, just because of agents, and, you know, for example, now documentation is king, I'm also interested in people like making different types of wallets as well, different, different form factors. For example, like Michael mentioned, someone But once again, it was super, super specialized and like you would really commit to that and bank on that as like a big thing, 'cause it would take many, many months or maybe even a year. Now something that someone can basically do on the side, because we've done the hard work of making sure the documentation, the APIs are as easy to use as possible." }, { "speaker": "michael_levin", "time": "48:36", "start": 2916.0, "text": "And, and importantly, it's self-custodial, right? So like, before you could build out stuff from a custodial perspective, but then once you start to get users, you have to go and do all the meatspace stuff, you have to get the licenses, you have to do all of that stuff, right? We've made it so that you can build self-custodially, but it's as easy as building, you know, from a custodial perspective, right? And I'm super excited to see what people, you know, are able to do with that, plus these AI" }, { "speaker": "olaoluwa_osuntokun", "time": "49:06", "start": 2946.0, "text": "Yep." }, { "speaker": "stephan", "time": "49:07", "start": 2947.0, "text": "Great." }, { "speaker": "olaoluwa_osuntokun", "time": "49:08", "start": 2948.0, "text": "This also allows us to like expand the markets. I think in the past, like many, many years ago, maybe we would go to a company and they're like, \"Oh, a Lightning node, I have to like manage this thing. Like, you know, am I doing fee updates? And maybe like, you know, they have to have, they have to have like a person that was sort of like specialized in managing these particular instances, and like there weren't many of those at the time, basically. This now we can just say, \"Hey, like, we'll take Or maybe like, you know, sudo host any component, it's still super powerful. If you look at all these, you know, vibe code apps that are coming out there or maybe platforms, they're hosting the database, the web server for you, and they're doing all that stuff, and that's still valuable in the current day, 'cause they can focus on like what's actually important for your application. In this case, we basically manage all the liquidity, we do the Lightning stuff and stuff, and this is about you sending and receiving payments and building top the application itself." }, { "speaker": "stephan", "time": "49:56", "start": 2996.0, "text": "Great. So yeah, I think that's a good spot to finish up. So listeners, you know, I encourage people to go and check it out, go and try vibe coding some of this stuff, see what you can build. you can go and find, Lightning labs at Lightning dot engineering, and, Roseph and Michael, will put the links to, for that so they can find you guys on X also. So thanks for joining me, guys." }, { "speaker": "olaoluwa_osuntokun", "time": "50:16", "start": 3016.0, "text": "Yeah, great. Thank you." }, { "speaker": "michael_levin", "time": "50:18", "start": 3018.0, "text": "Thanks for having us." } ] }