Latest / The Index Podcast / 🔥 Building the Safest DeFi Protocol on Solana — Plish, Co-Founder & CEO of Hylo
Transcript
- 0:00SPEAKER_01: I finally found a hardware wallet that isn't a pain in the ass.
- 0:04SPEAKER_01: Every hardware wallet I've tried, ledger, treasure, whatever, they all have the same problem.
- 0:08SPEAKER_01: They're annoying enough that you stop using them.
- 0:11SPEAKER_01: The Soulflare Shield is the size of a credit card.
- 0:14SPEAKER_01: No cables, no batteries.
- 0:15SPEAKER_01: You literally just tap it to sign transactions, and that's it.
- 0:19SPEAKER_01: It's built into Soulflare app, the same app that's been around since 2020, securing$14 billion for 4 million people.
- 0:25SPEAKER_01: And because it's Soulflare, you get real PL tracking, DeFi positions, and all your Solana stuff in one place.
- 0:31SPEAKER_01: Cold storage that you'll actually use.
- 0:34SPEAKER_01: Link in the bio, go take a look.
- 1:00SPEAKER_01: He's been building crypto since 2022 and previously held technical leadership roles at Blue Chip Silicon Valley companies, including Salesforce and SAP.
- 1:09SPEAKER_01: Hilo was founded in 2024 and won a prize in the Coliseum Radar Hackathon.
- 1:14SPEAKER_01: The company went on to raise a pre-seed and uh seed investment from Coliseum, Robot Ventures, and Solana Ventures.
- 1:21SPEAKER_01: And they crossed the 100 million TVL mark in late 2025.
- 1:26SPEAKER_01: Market conditions have changed, so it's not there anymore, but still significant traction for this team in a very short period of time.
- 1:32SPEAKER_01: Excited to have you on the show.
- 1:33SPEAKER_01: Thanks for joining me.
- 1:35SPEAKER_01: Thanks for having me, Alex.
- 1:37SPEAKER_01: And uh just to start off, I would love to hear more about your story.
- 1:41SPEAKER_01: You know, what gets you out of bed in the morning?
- 1:42SPEAKER_01: How did you get here and decide to build a Hilo?
- 1:45SPEAKER_00: Yeah, so I was uh doing actually crypto in the kind of corporate context uh before I started Hilo.
- 1:51SPEAKER_00: We were working on EVM chains and and doing some kind of enterprisey NFT and token stuff.
- 1:57SPEAKER_00: And um that kind of got me interested in the the smart contract programming uh thing.
- 2:03SPEAKER_00: And and the I guess what I ultimately realized is like it's not compatible.
- 2:07SPEAKER_00: Like the corporate enterprise world is not really compatible with Web3 or DeFi.
- 2:12SPEAKER_00: And so um decided to make the jump in 2024, um, really like late 23, like after watching um Tolie and the Dragon Suit at uh the Amsterdam breakpoint, uh got inspired to check out Solana, get away from EVM, and and start cranking on some um smart contracts on on Solana.
- 2:33SPEAKER_00: Um and what gets me out of the out of bed in the morning is like I would rather do this than anything else, even in these like awful market conditions and you know just the constant drama that this industry is in.
- 2:45SPEAKER_00: Um I would rather be in control of my destiny than like at a desk job being told what to do.
- 2:52SPEAKER_01: Yeah, I kind of feel the same way just generally about entrepreneurship uh and being an entrepreneur.
- 2:57SPEAKER_01: I just can't think of anything else that I could do or would want to do.
- 3:01SPEAKER_01: Um and even if I worked at like a big company, I'd still be an entrepreneur inside the company doing the same thing.
- 3:06SPEAKER_01: So, you know.
- 3:07SPEAKER_01: Yeah, I guess um yeah.
- 3:10SPEAKER_01: I uh tell me more about tell me more about Hilo.
- 3:12SPEAKER_01: What is it, what is it?
- 3:14SPEAKER_01: What have you guys been?
- 3:15SPEAKER_00: Yeah, so so Hilo is a is a DeFi platform on Solana.
- 3:20SPEAKER_00: Um the idea with it when we we started was pretty much to try to create something novel.
- 3:25SPEAKER_00: We found that there was a pretty big gap in the market at the time for a decentralized stablecoin.
- 3:30SPEAKER_00: And so we you know went through a whole bunch of different possible designs.
- 3:35SPEAKER_00: We looked at previous things that hadn't worked out, uh, namely like UXD, Hubble, a couple of like previous experiments uh in the decentralized stablecoin space.
- 3:45SPEAKER_00: And uh my co-founder had found this white paper from the ETH World with like a very interesting core equation.
- 3:53SPEAKER_00: And uh we said, all right, how can we uneath this and make it a cool product and bring it to Solana?
- 3:59SPEAKER_00: And um that's kind of where Hilo was born.
- 4:01SPEAKER_00: We we built a prototype, we pitched it to Coliseum, um, we won a prize and and kind of kicked off the journey from there.
- 4:10SPEAKER_00: Um and I can explain like in an abstract way what Hilo at its core does.
- 4:18SPEAKER_00: And we started with, I guess uh just to back up a little bit, we started with like sole LSTs as the the backing in Hilo V1.
- 4:26SPEAKER_00: And that's currently what's on mainnet, and that's currently what holds most of our TDL.
- 4:31SPEAKER_00: Um what the Hilo mechanism does is it takes a volatile yield-bearing collateral asset and it splits it into two possible trades.
- 4:41SPEAKER_00: People that want a stable coin that earns yield can do that, and then people that want to bet on the upside, like take the leveraged exposure, um, Hilo allows them to do that.
- 4:52SPEAKER_00: So we have this product called XSOL, which is uh kind of like a levered ETF on Seoul, and then we have high USD, which is a stable coin that can be staked for high yield, similar to like how um Ethena or whatever earn their yield.
- 5:06SPEAKER_00: The the cool part about Hilo is that it's all self-contained.
- 5:09SPEAKER_00: We don't use like a trading desk to place orders on Binance or whatever.
- 5:13SPEAKER_00: Everything is fully in the program.
- 5:15SPEAKER_00: We have no external dependencies.
- 5:17SPEAKER_00: We everything is just fully self-contained, which is um kind of a unique aspect in this current DeFi market.
- 5:24SPEAKER_01: Yeah, I mean, uh, this kind of gets in, I think I know where this is going here, but why those decisions were made.
- 5:30SPEAKER_01: And I think that kind of played out last week.
- 5:32SPEAKER_01: But I'll ask it this way: what is everyone else in the Salon of DeFi getting wrong right now?
- 5:38SPEAKER_01: That you guys have been right?
- 5:39SPEAKER_00: There's, I mean, obviously, composability is a big, big thing in DeFi.
- 5:44SPEAKER_00: This is the point, right?
- 5:44SPEAKER_00: You make um interesting financial instruments and you want other platforms to be able to use them, and things should be able to interconnect and collaborate and permissionlessly kind of compose um to make greater things.
- 5:58SPEAKER_00: But as an asset issuer, right, which is what Hilo is, we issue derivatives against Seoul.
- 6:03SPEAKER_00: We also issue a stable coin.
- 6:05SPEAKER_00: We don't really have a lot of room for plugging in other protocols.
- 6:09SPEAKER_00: We want to be plugged in, but we want to cap our risk at like just the assets underlying.
- 6:15SPEAKER_00: Right now we have Jito Sol and Hilo Sol as the kind of core collateral assets, and then we're gonna get into our V2 design later probably.
- 6:23SPEAKER_00: Um, but we're really kind of selective about what we take in.
- 6:26SPEAKER_00: In terms of dependencies, yeah, we we can't do that because as soon as you plug in another protocol and another venue, you inherit that entire company's risk profile.
- 6:36SPEAKER_00: Uh if it's like smart contract risk, so be it.
- 6:39SPEAKER_00: But if it's like OPSEC stuff, you also inherit that and like you're now liable for anything that happens to the operator behind that protocol.
- 6:47SPEAKER_01: And we saw this play out last week um with a couple of different protocols that were not, you know, with this drift, the drift hack, right?
- 6:55SPEAKER_01: Like some of these people who got wrecked were not actually drift.
- 7:00SPEAKER_01: There were other DeFi protocols that, you know, I heard of one story where they had um some kind of automated yield aggregate vault system or something.
- 7:10SPEAKER_01: And uh they their algorithm full ported like half the portfolio into this thing that looked like it had 80% yield APY, but it was just a uh, you know, it was the hacker with a fake token that that manipulated that pool, and that's part of how more money got sucked into that.
- 7:30SPEAKER_01: Um and so they they took advantage of what what not only obviously they compromised the multi-sig, that was step number one, but then they took advantage of how the system was designed to behave, which is that risk that I think a lot of people were not thinking about.
- 7:44SPEAKER_01: And it's interesting to me because there are lessons here we could learn from Ethereum from way back in like 2016 and even before then, like 20 since 2013, there's been multiple attacks like this.
- 7:58SPEAKER_01: Um, and so I don't know.
- 8:00SPEAKER_01: I I guess I'm kind of curious what you think, just calling it out bluntly, what's broken in DeFi that nobody wants to talk about right now.
- 8:08SPEAKER_00: Yeah.
- 8:09SPEAKER_00: And and that attack also is just like pretty classic DeFi manipulation.
- 8:13SPEAKER_00: You make a fake coin with uh a fake Oracle that you can kind of support with a bunch of on-chain watch trading to make it seem like it's worth something.
- 8:21SPEAKER_00: Um ultimately, like they they got the admin key and were able to register this thing as a market, which is something that only the the Drift Security Council was able to do before, right?
- 8:31SPEAKER_00: So like the admin key is is one whole topic, but um, being able to inject a fake coin and then borrow a bunch of money against it and rug all the liquidity out of the vault is what created that APY.
- 8:43SPEAKER_00: And so all these downstream consumers that were seeing this APY number, they just aped in, not realizing like, oh, maybe like the risk engine should be checking how much liquidity there is.
- 8:54SPEAKER_00: Um I just like dumping money in to be somebody else's exit liquidity.
- 8:59SPEAKER_00: Um so there's there's some lessons there in terms of like how to configure and allocate um resources kind of responsibly so that stuff like this doesn't happen again.
- 9:09SPEAKER_00: But to answer your core question, like I think the the core problem with with Solana DeFi these days is is just this reliance on on other protocols and the lack of novelty.
- 9:22SPEAKER_00: Like there aren't a lot of innovative yield sources.
- 9:25SPEAKER_00: Like we have the lending markets, they're great.
- 9:27SPEAKER_00: Those yields have compressed to like one, two percent.
- 9:30SPEAKER_00: And if you adjust for risk, it's negative.
- 9:32SPEAKER_00: Um, so we haven't seen a whole lot of like innovation here in in terms of what what you could possibly generate on-chain.
- 9:41SPEAKER_00: Like that old question of like where does the yield come from?
- 9:44SPEAKER_00: Most of these companies are pointing at, oh, those other guys, we integrated their lending pool.
- 9:49SPEAKER_00: Right.
- 9:49SPEAKER_00: And so like that's I think that's a that's a fundamental problem, and it's gonna inhibit the growth on this chain uh until we start seeing more kind of ETH style experimentation, DeFi summer style experimentation with new things rather than just like trying to build an app layer on top of somebody else's product.
- 10:07SPEAKER_01: Yeah, and I want to be really clear here too with my commentary and this whole thing.
- 10:10SPEAKER_01: Like, I'm not a security expert.
- 10:11SPEAKER_01: I only know what I've read on Twitter, and I'm not passing any judgment on drift and the guys that work over there.
- 10:17SPEAKER_01: They're super smart, hardworking.
- 10:19SPEAKER_01: You know, they built a really amazing protocol, and I feel terrible for them and for everybody who was impacted.
- 10:24SPEAKER_01: And I know a lot of people who were, who lost significant sums of money from that.
- 10:27SPEAKER_01: And it's just it was a terrible it was a terrible week uh for the whole ecosystem, but one that we're gonna recover and learn from.
- 10:34SPEAKER_01: Um, and you know, I just wish I wish everybody the best getting through it, and hopefully law enforcement catches up.
- 10:39SPEAKER_01: And I think this is one of the things that's come up for me.
- 10:41SPEAKER_01: You know, I've been podcasting for eight years, and my first show was about the intersection of privacy, technology, human rights, and democracy.
- 10:49SPEAKER_01: And um one thing I learned from talking to lots of experts, including, you know, Cindy, who's the director of the EFF, um, the guys behind Zcash, and a bunch of other you know, people who are experts in privacy, for example.
- 11:04SPEAKER_01: Number one, what happened wasn't what there were like tornado cash was used as part of this, I think.
- 11:09SPEAKER_01: Um, but even then, like law enforcement has lots of means and methods for tracking and tracing and figuring out who was responsible.
- 11:16SPEAKER_01: And like early indicators obviously point towards a state actor, and you know, you got to be super hardcore with your OPSEC, but also like it's you know, you're going up against really sophisticated adversaries, and sometimes you just you just lose, I guess.
- 11:31SPEAKER_01: Um but um anyway, I think from what I've seen from the foundation, they've really leaned in and uh stepped up to support teams and provide like an extra layer of guidance and security and and and and things like that.
- 11:44SPEAKER_01: So I think that's gonna help help a ton.
- 11:47SPEAKER_01: Um I'm curious on the innovation side, what's uh what's Hilo Hilo doing right there, you know, that everyone else isn't getting?
- 11:57SPEAKER_00: Yeah, I think uh the the main thing that we did right was we introduced a novel concept.
- 12:04SPEAKER_00: And to give some color on what it's like to bring a novel concept to this market, um, it was really hard to raise venture in the early days uh because it was kind of an untested mechanism.
- 12:16SPEAKER_00: We had, you know, kind of a unique, almost esoteric alien type of thing that we brought to the table where everything else was kind of standardizing around like normal stablecoin yields at the time.
- 12:28SPEAKER_00: So a lot of ECs wouldn't touch what we did.
- 12:30SPEAKER_00: So we were like very thankful for Coliseum and for Robot and Solana Ventures for believing in something new coming to the chain.
- 12:40SPEAKER_00: Um so yeah, what what I think we've done right is essentially do the hard thing over and over.
- 12:46SPEAKER_00: Um we we had many options, including in our current V2 roadmap, where we could have plugged in other stuff.
- 12:54SPEAKER_00: Um, we could have been a little riskier, we could have been a little bit more like narrative chasey, but we kind of stuck to the most kind of grounded conservative truth that we knew about.
- 13:05SPEAKER_00: And like our core yield source for our stablecoin at this point is the Solana LST uh ecosystem, right?
- 13:11SPEAKER_00: That's been what the V2, the V1 has been built on.
- 13:14SPEAKER_00: And so we knew that that's multiple times audited and verified, and it's like a really, really simple program.
- 13:21SPEAKER_00: So, like, sure, this is a great thing that earns sole on sole yield.
- 13:24SPEAKER_00: How do you transmute this into dollar yield?
- 13:27SPEAKER_00: And um, and how do you hedge all the risk under underneath it?
- 13:31SPEAKER_00: And so what we did, I think, right is like build off of like really good uh bricks.
- 13:36SPEAKER_00: And then we also did our risk management really well.
- 13:39SPEAKER_00: So our V1 risk management has gone through the 1010 crash, the February crash, even this incident with really not much, right?
- 13:48SPEAKER_00: Like it's it's acted exactly as it's supposed to act.
- 13:51SPEAKER_00: Um obviously there are like downsides to how it acts, it's very aggressive, and so we've had to lose TVL as a result, et cetera.
- 13:59SPEAKER_00: But none of that TVL is like leaking, it's purely just like de-risking how much stablecoin is in the system because the prices are going down.
- 14:07SPEAKER_00: And obviously, we learned a lot of lessons there, and we're gonna expand into more collateral types so we're not just tied to Solana's fate, uh the L1 token fate.
- 14:17SPEAKER_00: But that's kind of what we did, is we we just did the hard thing and and we manage risk really well programmatically.
- 14:22SPEAKER_01: Yeah, I think that makes a lot of sense to use the vetted RE open source contracts.
- 14:27SPEAKER_01: I think that's one thing that I would really like to send a message to Solana DeFi founders out there is like everybody should be open sourcing their stuff.
- 14:35SPEAKER_01: I think that that is actually way safer, and people people have this feeling that there's a moat around the source and there's just not, right?
- 14:42SPEAKER_01: Especially in the age of agentic commerce and agentic the agenc web.
- 14:47SPEAKER_01: Um agents can just recreate your functionality pretty quickly.
- 14:52SPEAKER_01: Um and so you know the moat really is around the network effects around that TVL and the other and the yield you can create and the partnerships you can do.
- 15:02SPEAKER_01: Um but I I think I think open sourcing a lot of this stuff lends to better security.
- 15:08SPEAKER_01: Um that wasn't the case for what happened last week, but it has been in the past.
- 15:13SPEAKER_01: Um and you know, getting more open source battle-tested contracts out there that people can compose with is I think gonna improve everything.
- 15:22SPEAKER_01: Um I'm curious, have you checked out what the Solomon Labs team is doing and their stable coin?
- 15:30SPEAKER_01: Um, I've had them on the show, respect those guys.
- 15:32SPEAKER_01: Curious what your opinion is on their their that asset and you know, generally their their their model.
- 15:38SPEAKER_01: Because I do I think they they share some philosophies um around building their own stuff that like you guys have.
- 15:45SPEAKER_01: Um so curious what you think about that.
- 15:47SPEAKER_00: Yeah, yeah, sure.
- 15:48SPEAKER_00: I'm I'm somewhat aware with what they're doing.
- 15:50SPEAKER_00: I mean, if I understand correctly, they're running kind of basis strategies, but the the the way that they deliver the yield is unique.
- 15:58SPEAKER_00: You don't need to be in their program to earn the yield.
- 16:02SPEAKER_00: So if that stablecoin is deployed somewhere like in a AM or some other type of pool on chain, they have uh kind of a back-end system that sends the yield to to where it's at.
- 16:13SPEAKER_00: Uh I think that's pretty cool.
- 16:15SPEAKER_00: It's it's you know, I wouldn't say it's like very it doesn't fully align with how we work.
- 16:21SPEAKER_00: Like we want everything to be on-chain in their case, like you're relying on a back-end system to do the work.
- 16:26SPEAKER_00: There really isn't a way to distribute the yield uh programmatically that way on chain.
- 16:31SPEAKER_00: Um but I think it's cool, and I think we need more things like this uh regardless, like more alternative stable coins, more yield sources that are kind of unique and and done correctly.
- 16:44SPEAKER_01: I think my my thing with all of these yield generating opportunities with my assets, like the safest thing to do is to have a cold wallet, a ledger that doesn't touch anything except for your soul and uh and stakes it.
- 16:59SPEAKER_01: That's it.
- 17:00SPEAKER_01: It can stake and that and unstake and it can send rewards to another wallet that you use to liquidate if that's what you want.
- 17:06SPEAKER_01: Um, but to me that's the that's the only thing that I that I that I consider doing because it just feels the safest.
- 17:15SPEAKER_01: Like, dude, I mean I had I had people recommending to me not seven months, seven months ago or something, like, hey, you could just do this thing and and it involved drift, and I probably would have gotten rugged, you know, um, you know, if I had done it on a contract where I don't retain custody.
- 17:29SPEAKER_01: So I think the self-custody and being able to control that in the in Liquid Soul also works that way, right?
- 17:35SPEAKER_01: It's still self-custodied.
- 17:37SPEAKER_01: Um, I think where you get into trouble is when you're not custodying your own assets.
- 17:41SPEAKER_01: Um that's that's where there's that's where there's some risk.
- 17:45SPEAKER_00: Yeah.
- 17:46unknown: Yeah.
- 17:46SPEAKER_00: There's a lot of also kind of like there's a lot of crazy strategies that you see, which are cool, right?
- 17:52SPEAKER_00: Like you'll see these these ex posts of uh there's a lot of these yield farmer kind of thread poster guys that come up with like, ah, you can deposit this thing over here, you borrow borrow it over here, put it into like this yield splitter, and then take the yield splitter thing and deposit it over here and loop it.
- 18:07SPEAKER_00: Right.
- 18:07SPEAKER_00: And that's cool.
- 18:08SPEAKER_00: I think it's very creative.
- 18:09SPEAKER_00: Like I really do like that people come up with this stuff, but like if you're gonna full port a strategy like that, um, you have to observe like how many layers of risk am I inheriting by by crossing three different protocols uh to just like burn people to do that.
- 18:24SPEAKER_01: Think about it, right?
- 18:25SPEAKER_01: Like you've got you've got this risk pyramid where where it's like um you know, at the top is like the riskiest part of your asset allocation, you know, so it should be a very small uh amount of relative capital to what you have that you're doing stuff like this with.
- 18:40SPEAKER_01: Like it's it's crazy the compounded risk over protocols, like, you know, especially when the software's not open source and you can't audit it yourself.
- 18:47SPEAKER_01: Like right now, what's awesome about today that's way different from you know when I first got into this industry, I've got AI.
- 18:55SPEAKER_01: I've got a my team, we call them clankers.
- 18:57SPEAKER_01: I've got a clanker that I can just like drop a repo into and run a code review on, and it can tell me right away how much if there's any security vulnerabilities.
- 19:05SPEAKER_01: So that's not perfect, right?
- 19:07SPEAKER_01: Like it doesn't, it's not perfect, but it's better.
- 19:09SPEAKER_01: You know, I don't have to hire like a full-time security engineer to help me manage my risk across from a software from a contract standpoint, if you can read what's going on under the hood.
- 19:21SPEAKER_01: Um, so and actually I'm thinking now too of my CTO, like my CTO and one of our engineers had this idea for a piece of software to uh basically be able to, even if it wasn't open source, basically like still reverse engineer like what's actually happening under the hood.
- 19:38SPEAKER_01: And now I'm realizing that's probably a pretty good idea for security purposes.
- 19:42SPEAKER_00: Yeah, it goes the other way, you know?
- 19:45SPEAKER_00: Yeah, it totally goes the other way.
- 19:46SPEAKER_00: Like uh there is a potential downside to open sourcing, at least like the account surface area of a program, because like if you're a newer team, hasn't been audited that many times, the attackers are using really good prompts to try to like find where there are weaknesses in your system.
- 20:05SPEAKER_00: Um, and there's also like this added kind of thing that that a lot of these teams do, which is let other teams into their code base, even if it's closed source.
- 20:13SPEAKER_00: And we've seen some exploits come out of that where like you trusted somebody even under NDA, and somehow that code got into somebody else's hands.
- 20:20SPEAKER_00: And and these days, within you know 15 minutes, Claude can find a vulnerability and shape like an attack plan uh for a bad actor.
- 20:28SPEAKER_00: So it cuts both ways, like this technology.
- 20:31SPEAKER_01: Yeah, and it's crazy with what um I did you see that like that leak?
- 20:35SPEAKER_01: I'm using air quotes here for people listening from anthropic where they leaked mithril or whatever that that new model is that's like trying to break out and can find like found like all these security bugs all over the internet.
- 20:48SPEAKER_01: I mean, it's it's getting a little it's getting a little dicey, I think.
- 20:53SPEAKER_00: Yeah, a little maybe a little too powerful, but I want I want that, right?
- 20:57SPEAKER_00: I want access to that so I for our own good to find stuff that our auditors and us and our tests and all of the stuff that we do doesn't catch and let's slip through.
- 21:07SPEAKER_00: Like it's it sounds like an amazing tool, but yeah, again, it's double-edged.
- 21:11SPEAKER_01: So um I would be really curious from your time in corporate, you know, the corporate, more corporate corporate world, both in terms of crypto and AI, like how do they how are they viewing this space?
- 21:25SPEAKER_01: Like, do you still keep in touch with folks that you used to work with over there?
- 21:27SPEAKER_01: And like what are they doing with agents?
- 21:29SPEAKER_01: And are they talking about agentic commerce at all and wanting to tap into that space?
- 21:34SPEAKER_01: Just really curious what you're hearing.
- 21:37SPEAKER_00: Yeah, I mean on the AI side, I actually worked on AI um at Salesforce.
- 21:42SPEAKER_00: Like we there was a thing called Einstein that they were pushing for a long time.
- 21:45SPEAKER_00: It was like their embedded embedded AI in all of their software systems.
- 21:49SPEAKER_00: So like I worked on pieces of that.
- 21:51SPEAKER_00: Um, so we were early on like actually transformer models.
- 21:54SPEAKER_00: Like we had uh TensorFlow in production, like flagging emails for having like an executive.
- 22:00SPEAKER_00: Attached or whatever, like really, really primitive use cases where you had to train the model to with a bunch of examples and and like encode what you wanted.
- 22:08SPEAKER_00: Um, I think this is you know essentially opened up like a big space for like startups to just jump in and do it better than corporate, because corporate has a lot of bloat and they want to tend to own things.
- 22:21SPEAKER_00: And there's also a lot of rules around like owning the data where you have like an MSA with some some company that's buying your products and you have access to their data, that may not involve training, right?
- 22:34SPEAKER_00: And so there's a lot we found a lot of limitations at that time around um what you can even do with the data.
- 22:39SPEAKER_00: But from what I hear, yeah, there's they've fully leaned in to some form of agentic coding at the very least.
- 22:47SPEAKER_00: And there's a lot of interesting use cases coming out of these kind of major enterprise companies at this point for their for their customers, but it's never as bleeding edge as what we see and what we're doing in our company.
- 22:59SPEAKER_01: Yeah, what are you guys doing?
- 23:02SPEAKER_00: Yeah, we I mean we use uh a lot of agents.
- 23:04SPEAKER_00: We we only have a team of four, right?
- 23:06SPEAKER_00: So we're extremely lean and uh we've we were three before, we're now we're four.
- 23:12SPEAKER_00: We're not really planning to like scale up too much because it's too powerful.
- 23:16SPEAKER_00: Like I literally had a an instance today where you know we were looking at um potentially bringing somebody on as like a contractor to help with um building out like a piece of of infra.
- 23:27SPEAKER_00: And I asked my co-founder, like, hey, like maybe this guy could help you do X, Y, and Z.
- 23:30SPEAKER_00: He's like, Oh, I already did that.
- 23:32SPEAKER_00: Right?
- 23:32SPEAKER_00: And and so it's the the multiplier on our productivity just with like having agents in the loop and constantly scanning our own code bases and doing audits for us before we pass it to the auditors, has really increased the the productivity.
- 23:48SPEAKER_00: Um, where we haven't really stepped too far in is like having the agents analyze meetings and stuff.
- 23:53SPEAKER_00: I don't I don't really like the tools that that exist currently, so it'd be more fun to maybe build our own for like the kind of the human side of what we do.
- 24:02SPEAKER_00: But in terms of code and and code review and code production, we're so fast now.
- 24:07SPEAKER_00: Um it's like our v2 timeline is shrunk in half.
- 24:11SPEAKER_01: I'm I'm using it a ton uh on the human stuff.
- 24:15SPEAKER_01: Um with uh both with hiring, yeah, with hiring and with customer calls.
- 24:19SPEAKER_01: Um we use granola uh to do to like record everything.
- 24:24SPEAKER_01: And then um we have a workflow where we've got some skills that we built into Claude and um like a Google Drive, and we have like this whole hiring process.
- 24:33SPEAKER_01: I can share it with you.
- 24:34SPEAKER_01: It's based on this book called Who, um, and it's probably the best hiring process I've ever used.
- 24:39SPEAKER_01: And we have built out a system on Google Drive with like templates for each stage in the interview and like scorecards for the each candidate.
- 24:49SPEAKER_01: And then, you know, I use Quad to like take the transcripts and take the notes from granola, put it into the interview based on which stage we're at, and then you can just like talk to your agent because it knows the process that you're going through, the whole who process, because we've got a skill for it.
- 25:04SPEAKER_01: Um, and it helps me think through what I heard versus like objective.
- 25:09SPEAKER_01: Like I find it's way more objective than I am about, hey, does this person meet the requirements for what we're hiring for?
- 25:16SPEAKER_01: And it kind of it keeps track of the patterns across interviews, and then the next person, the process that needs interviews is the same workflow.
- 25:23SPEAKER_01: It all gets updated, and then when we're, you know, with these interview, we do like seven or eight interviews for every candidate, which is seems crazy, but it's super important in a startup, like get the right people.
- 25:32SPEAKER_01: Um, and at the end, you know, we're able to get a really full picture of the patterns across all these interviews about this candidate and really understand like, is this person an A player and are they a cultural fit for us?
- 25:43SPEAKER_01: Um, and it's been working great.
- 25:45SPEAKER_01: It's been working great.
- 25:47SPEAKER_00: Really cool.
- 25:48SPEAKER_00: Yeah, you can get a lot of data out of that process that probably was missed in old interview processes.
- 25:54SPEAKER_01: Oh, dude, totally.
- 25:55SPEAKER_01: I mean, I I definitely I've gotten some wins from it so far.
- 25:59SPEAKER_01: Um, it's been really interesting.
- 26:01SPEAKER_01: So, what's what's um back to Hilo though, what's next for you guys in the next like couple months?
- 26:05SPEAKER_01: What are you most excited about?
- 26:08SPEAKER_00: Yeah, a lot.
- 26:09SPEAKER_00: I mean, we we have uh learned a lot in this kind of bear market um from like our V1 and kind of the shortcomings of just like being attached to one collateral asset.
- 26:19SPEAKER_00: And so, you know, we listened to the customers, we we looked at what happened on 1010, we looked at how our protocol behaved, and you know, obviously everything worked as intended, but there were a lot of like side effects that we didn't like, right?
- 26:32SPEAKER_00: Nobody wants to lose TDL, right?
- 26:34SPEAKER_00: So we wanted to come up with a more robust system with better risk management, better scalability, and way more options for for people to trade with.
- 26:44SPEAKER_00: And so what we've been working on is is twofold, right?
- 26:46SPEAKER_00: We have the the v2 UI, which is I think coming out maybe this week or by the time um this podcast comes out.
- 26:53SPEAKER_00: So we had a bunch of early feedback from our like inside circle of users and then a lot of public users as well.
- 27:00SPEAKER_00: And so the new the new UI is like a big uh step up from where the v1 was.
- 27:06SPEAKER_00: Um so it's a lot more kind of perp trading like for the for the X asset tokens like XSOL.
- 27:13SPEAKER_00: Um there we're adding in kind of hooks for uh stop losses and DCA and a whole lot of stuff coming down the pipe so that people can have more tools.
- 27:23SPEAKER_00: And there's a lot more like risk readout stuff that they can look at now, like break-even prices for XOL, um, and and a lot better like visibility around what happens when risk management intervenes.
- 27:36SPEAKER_00: So the UI is kind of one whole uh angle of this that we learned a lot from and improved.
- 27:41SPEAKER_00: But uh what I'm the most excited for actually is the protocol V2, which has uh been audited once already.
- 27:47SPEAKER_00: We have some things to crank on internally and then probably do a second audit before it comes out.
- 27:53SPEAKER_00: And so the V2 protocol moves us into a world where we have many different types of collateral assets.
- 27:59SPEAKER_00: So we're no longer tied to just sole LSTs.
- 28:02SPEAKER_00: Obviously, those will still be like bread and butter for what we do.
- 28:05SPEAKER_00: But the first thing we're gonna be launching on V2 platform is XBTC.
- 28:09SPEAKER_00: So bringing in a wrapped Bitcoin uh collateral, creating this levered uh 3x ETF on BTC, and then also just further cementing the possibilities for like having strength in the stable coin so that like even though Sol and Bitcoin are are correlated, um, the volatility on Bitcoin is a lot lower.
- 28:30SPEAKER_00: So it'll kind of stay more stable in a sense.
- 28:33SPEAKER_00: Uh and it goes beyond Bitcoin, right?
- 28:35SPEAKER_00: So this is kind of this is gonna be the first kind of test rocket uh out of the the V2 protocol, but we're also looking at you know Zcash, we're looking at uh other yield-bearing dollars, we're looking at different RWA plays that we can bring in.
- 28:48SPEAKER_00: We could consider even stocks, like um imagine like a stablecoin backed by uh the SP 500 index, right?
- 28:56SPEAKER_00: So so we want to bring in this kind of breadth of assets and commodities and really anything that can be tokenized and that has like staying power in the market.
- 29:05SPEAKER_00: We're trying to just broaden the base and become this kind of giant autonomous hedge fund like thing.
- 29:12SPEAKER_01: Very cool, super exciting.
- 29:14SPEAKER_01: So um, we're getting to the top of the show, and I always ask this, but what have I not asked you that I should have asked or that you wanted to talk about?
- 29:21SPEAKER_00: Um I I think like leverage and and like the perps war maybe is is a piece that we can get into a little bit.
- 29:29SPEAKER_00: Um other than that, I don't know, like user education.
- 29:35SPEAKER_00: There there's there's a few things that tell me about it.
- 29:38SPEAKER_01: I know nothing about leverage or the perps, the perps world at all because I don't touch it.
- 29:43SPEAKER_01: I think you probably can tell from my conversation here that like I'm extremely low risk when it comes to this stuff.
- 29:48SPEAKER_01: Um I also just don't have the time in the day, man, to like manage that.
- 29:52SPEAKER_01: You know, it's too scary to me.
- 29:54SPEAKER_00: Yeah, yeah.
- 29:55SPEAKER_00: I'll get on my soapbox about leverage a little bit, I think.
- 29:58SPEAKER_00: Um and that's just specifically about the DeFi landscape and how it makes leverage very accessible and almost like a game, right?
- 30:08SPEAKER_00: You're seeing this with like the normal perps thing, but you're also seeing this with um looping strategies where people are able to take like 10x loans against collateral and buy more and and um essentially do stuff that was guarded by you know trading desks with prime brokers for a long time or like FX desks.
- 30:27SPEAKER_00: You know, they're taking giant loans at really, really high leverage to like make these kind of cross-market uh carry trades and stuff.
- 30:36SPEAKER_00: Um what DeFi has done is democratized all of this, but you know, the the audience is not necessarily like fully capable of assessing risk when it comes to these things.
- 30:46SPEAKER_00: So like you can open a perfum and hit the 50x button and put in five bucks, 10 bucks, 20 bucks, whatever, but like the risk is so high, the sensitivity to delta is so high that like you can just immediately take a haircut if the market moves.
- 31:02SPEAKER_00: One tick later, the market moves down, like you lose all your money.
- 31:05SPEAKER_00: Right.
- 31:05SPEAKER_00: And so our our stance on this, especially as like Solana is going through this identity crisis around how do we beat hyperliquid?
- 31:14SPEAKER_00: How do we make a perps market that's um that's better and can like command a bunch of user attention?
- 31:20SPEAKER_00: Our stance on this is maybe you don't need as much leverage.
- 31:23SPEAKER_00: Maybe you should be a little conservative.
- 31:25SPEAKER_00: Maybe you should use a product that's like managed, and that's kind of what the X assets do in our platform.
- 31:31SPEAKER_00: They're right now capped at like two to four X, ideally three X, and the risk management intervenes to delever the trade as you're directionally wrong.
- 31:42SPEAKER_00: And I think that's like a really important thing, especially for a nascent user base, which we have on crypto and DeFi in general.
- 31:49SPEAKER_00: Like not everybody's a pro.
- 31:51SPEAKER_00: Uh, we think it's a really, really powerful unlock, and and we think that like the perps war is is great, and I hope we win, but we're providing an alternative that is entirely kind of separate and and a lot um more contained in its like risk surface area with X assets.
- 32:08SPEAKER_00: Awesome, very cool.
- 32:09SPEAKER_01: Well, thanks for coming on the show.
- 32:10SPEAKER_01: It's been good to have you here and hope to hear from you again soon.
- 32:13SPEAKER_01: Thanks for having me on, Alex.