Why Crypto Still Struggles to Capture the Value It Creates | Roundup

Transcript

You are seeing this trend in crypto right now at the open source layer at, you know, at the credibly neutral layer of actually making certain parts of that more closed source over time to capture more and more value. And I would make the argument that Ethereum actually is still potentially the only one with the north star to be a purely open source. Quick shout out to today's sponsor, Katana. We'll hear more about Katana's 1 billion cat campaign later today, but for now, let's get into the episode. Hey everyone, quick disclaimer before we get into today's episode. Nothing said on bell curve is a recommendation to buy or sell securities or tokens. This podcast is forformational purposes only and the views expressed by anyone on the show are solely our opinions, not financial advice. Our guests and I may hold

positions in the companies, funds, or projects discussed. All right, everyone. Welcome back to another episode of Bell Curve. You got me, Miles Zave. Happy New Year. How's the time off? >> Great. Great. Happy New Year. Got some sun in Florida. Are we past the Larry David deadline for saying happy new year? I think that's Jan 3 based off of Larry David's uh u evaluation. >> I'm actually Is the Is the rule just you can't say happy new year after a certain period of time. That feels like a Larry David rule. >> Yeah. Yeah. Exactly. Exactly. We're getting close.

>> Miles, you had some heart attacks over the the break. Uh some >> I did. I did. I did. a a vicious 20 pound terrier bite my fiance. Uh so it's it's it's just he's getting old. Uh so yeah, don't >> he was threatened. You have a new lady in your life and he was threatened. He's like, >> I guess my miles. >> That's his way of showing it. But yeah, apart from that, all good. A year is off to a great start.

C you know what it certainly is. I think 2026 is going to be a good year and we we are going to speak today mostly about software. So there is a the theme to this episode is commercial open- source software. So we're going to go back like really brief history of just how software came to be as an industry because similarly to crypto it was kind of this amorphous thing that there was no IP protection around no business model around and you know it kind of grew in this way that disrupted the the profit stream of the computing industry which was very hardware oriented uh but it did give rise to open-source software which people didn't really understand how to make money on for a long time. uh you know you have some success cases like Red Hat it's used as a tool within

big web 2 um competition you know the Amazon's and the Googles of the world have leveraged open source as a as a tool as it were uh but L1's fundamentally broke this trade-off and so we want to introduce crypto into this story and lineage lineage of software and then I would love to actually end on AI and how AI is even complicating the picture further and that'll lead into this uh kind of three-part series that the three of us are going to embark on on AI and crypto and what the intersection looks like there because even though it went through a little bit of a hype cycle earlier this year and last year or 20 in 2025 and late 2024, I think we're all convinced that there's a there there and it's very interesting.

And so we're going to do a deep dive on that and maybe sift through the wreckage of some of the broken ideas and uh try to find some diamonds in the rough. But uh Miles, I know you've done a bunch of research on this. maybe I can just kick us off with the just to kind of set the context of the history of software in general um and how it evolved and then maybe I could turn it over to you for the development of open-source software and some of the early business models but it's a really interesting history and I've talked about this on the podcast before I would highly recommend the book how the internet happened by Brian McCulla who talks about some of the early days of the internet software et now he kind of starts with the internet but the the through line in

between the development of the internet and software in general is that it was largely academic and in many cases government and kind of researchfunded right so for the internet you had arponet which was part of the defense apparatus of the United States in the early days of kind of computing you know the business model was entirely about selling large computing mainframes right this was the IBM business model you'd sell a computer the size of a 18-wheeler you know to some university or government um you know government entity or something like that and the software was really an afterthought and you did have some early uh pieces of software like Unix coming out of Bell Labs but it was really pretty sleepy until I think

really Microsoft was the the first massive company um that got built that showed that you actually not only could uh make money with software but that you could actually be the platform and extract and disrupt a lot of the the hardware profits and that was kind of the core platform that was super super valuable and some there's some really interesting stories of those early days of you know IBM this sounds hilarious to say now IBM just didn't think software was coarse so they licensed out Microsoft to build software the operating system as it were for a bunch of the early machines and you know Microsoft didn't actually build it themselves they they licensed DOS which was kind of this open source thing for uh for $50,000 I mean there are these crazy stories um you know that QOS

license is probably the best uh the highest ROI investment in the history of business and even in those early days the business model of closed source software was a huge question mark because uh software was not legally uh intellectually protected IP. It wasn't IP so you had no protections. Anyone could just go in and copy it. Microsoft was part of the lobbying effort that uh you know branded software as um intellectual property. However, what that did do was imagine, you know, as nerdy as crypto is, imagine the early internet enthusiasts that were building software when computers didn't even had screens. They just had these little lights, you know, that you would input your software or it would output a certain series of lights like Morse code uh that did a calculation for you.

Imagine how those types of people. And so when Microsoft and some of the early software businesses pushed to do closed source and treat software as intellectual property, it led to this movement that would eventually become open-source software. So you just had people similarly this will probably rhyme with a lot of the folks that you see in crypto today which are who are highly ideologically driven and just said this there's just an ideological perspective that code and writing code should be free and that cultural idea uh that meme ended up being really really powerful and some of the early open-source uh you know tools were uh there was kind of GNU and free software

so there's a guy Richard Stallman who launched the uh genu project in 1983. Um and he really was kind of this software is a it's a freedom and it's a moral issue kind of guy. The issue is that there were a bunch of compilers and tools and libraries but there was no kernel. Uh that got introduced in 1991 with Linux tovaults. So he was a Finnish student um and he just had this hobby kernel called Linux. And the the the magic ended up being the combination of these tools, these libraries, these compiler, these free open- source tools with the kernel. And that gave birth to the Linux operating system. And you know, there's a whole there's a whole really interesting history here that probably doesn't bear mentioning. Um,

but a lot of things happened like you might see in crypto today, right? Because Linux wasn't a for-profit entity, they had to figure out things like governance. And the other thing that ended up happening was this grassroots community of developers that contributed for free because of this ideological buyin and this um you know resonating with the message of moral freedom and things like that and Linux Linux really grew and developed and became highly used. The question was how do how do you monetize it? And people didn't make for for some and this proceed this is kind of the origin of many of the value creation versus value capture stories that people love to talk about in web 2 and it found their way into web 3. Um I don't know I'm monologuing it here a lot miles I don't

know if you want to like Red Hat was a super interesting development in this history and that gave birth to this investing style and this whole category of commercial open source. So, I've talked way too much already. Can you take us away with the red? >> That's perfect. Perfect leadened. Um, yeah, I think there's a lot of things that rhyme between like uh this open core commercial open source software space and crypto. Um, and I think in general we're doing this uh episode. um because there are lessons we can draw from I think the earlier you know open source and like open core uh because it came earlier and there's just like a lot of things that rhyme with what we see in crypto um and so you know I think OSS

was like you know kind of as you said like a push back against making you know everything proprietary and seeing these like large incumbents just you know kind of cement their advantages right um And it really started with I think which is even older which was like the free software movement, right? So there's like an ideological movement that software should be free. Um and frankly it sounds a lot like the early like crypto folks. It sounds like a lot like uh the early Bitcoin core devs and community as well as Ethereum, right? Like we're talking about open access to services um and those services are run on software. And so, but maybe backing up a little bit to to open core like um there was a split in the communities

between these ideological folks that believe everything should be free um and folks that looked at open source software and sawm this is an interesting com like commercial strategy, right? And so how this usually works is that somebody will put out some sort of open-source software product or service. Um, and in like a very bottoms up grassroots motion. Um, this will become, you know, if it's good, it will quickly become loved within like developer communities, right? This is like what they're playing around with on the weekends. Uh, you build up a ton of GitHub stars. Uh word of mouth is like very very powerful within these

communities. You'll see the stuff on like hacker news, right? Um and then once you have like a uh you basically use open source as like a distribution hack, right? Um and from there you commercialize or monetize your open source project through something that's called like an open core company. And you know crypto like analog is like this is like the development companies right associated with this is like the red hat for Linux right when the Linux foundation is maybe like the you know like token foundation that we see in crypto a lot um and it's usually centered around like adding proprietary features and reducing friction generally

right so like they will offer a hosted version of the SAS uh like a SAS version or a hosted version of the open source software to reduce like DevOps headach they will think about okay what are like proprietary features that we should only include in the paid version even if we maybe make all like the source code available like some of it's going to be MIT license some of it's going to be like proprietary right um again we see that in crypto a lot and I think like the the general sort of philosophy has been um you keep like best in terms of best practices like the features that are used by individual contributors um those are most similar to your open original open source users those should be part of the open source codebase. The

features that you know um buyers or managers of the software want. So this is things like role based access control or like maybe even just like usage limits like that should go into the enterprise you know SAS version that you actually pay for. Um, and maybe what I'll say like on the open core side is that there are like it is it's not a new category. It's been around since Red Hat. Um, it's definitely grown a ton as a category, but there are still a lot of skeptics. Um, and those skeptics like the conversations between like the proponents of open core companies and the skeptics of it sound a lot like what we hear in crypto where you know I think

like the at some point if you're not careful you run into a a tricky trade-off space um as a company that is like both the proprietary like or like the primary contributors of the open source project and the ones trying to monetize that project with a company. Um, you know, at some point you have to kind of decide like should we if we have 20 to 50 people on the team, how many people should be focused on adding features to the open source version versus the paid version, right? And what we've seen quite a bit is that you obviously like the incentives are such that you want to prioritize the paid features. Um, and so we've seen the pitfall in the past of like open source

communities losing uh or open source projects losing the communities that made them successful because you know they just tried to monetize um as much as possible and they basically stop investing and making the open source project like you know competitive right um or you know you relinquish your control of the open source project uh governance comes into play like the only time I've ever heard heard like software governance outside of crypto is in the open core communities. Um, and they don't usually have tokens, right? So, it's actually like kind of complicated.

Um, but we've seen really successful versions of this, right? Like Supabase is like a a great great example of this. So, they took, you know, they built like a database project on Postgress SQL. Um, >> and people love it. people like really love it and it's now like a legit competitor to you know large large companies like Google right for for database software. Um, so yeah, maybe I'll pause there just on like that's the open core side of how like where does open source software like touch in a commercial setting like the first one is this rise of open core companies and this is like a a really cool category frankly I find it interesting to study um recommend folks checking out like open core ventures or um OSS capital uh

those are two like venture venture companies that have like uh wrote a lot about this this space and I There's a lot to learn, I think, uh, as we'll get into. >> Yeah. So, I think that maybe to start to parse these out into different eras. I think the Red Hat story was a really big one because it proved to investors that there was a way to so you kind of had this trade-off, right, of closed source software which was, you know, highly scalable, very monetizable, uh, and open- source where it had these superpowers where you actually had a bunch of contributors who were highly in many cases highly, highly talented contributors who would work for free and get wide adoption, but as soon as you tried to monetize it at all, you lost

the core advantage. And so it actually there was a period of time in tech where it felt like both sides were going towards their extremes and stagnating. So you had the the Balmer era of Microsoft where it was just you know enterprise software and kind of but kind of boomer central and they were losing talent. They couldn't get developers or anything. At the same time you had Linux which was taking off but you couldn't monetize it. And Red Hat was this super interesting data point that even if you don't sell the software, you can sell things like stability and build a big business that they eventually sold to IBM for somewhere around the tune of $34 billion. And that that was a strong data point. And I think since then, a couple of things have happened. We're going to end this episode on crypto and AI, which

are huge developments, frankly, in this long tapestry and trajectory of open source software development. But two other things that happened were uh cloud. Um, cloud was a huge moment in open-source software and it made web 2 giants get far more soph sophisticated about how they leveraged open- source software as a competitive tool within their tool belt um often to neutralize the advantage of uh people that they were competing against. There's a great example here of uh Kubernetes uh being released by Google to commoditize what something that a product that Amazon was trying to release. I also think maybe we don't really need to talk about this, but I do think GitHub and this kind of era that we're entering of developers uh

producing code on Amazon for like kind of status and clout and cred to leverage job offers and things like that later. Crypto, there are so many things like that that rhyme with crypto and CT. Um I I do think crypto is really good at anticipating the future, often unintentionally, but I think that's another thing interesting thing to call out. I I do want to call out um I I want to go into this era of cloud and maybe we can bleed a little bit into AI here about how big corporates think about leveraging open source software as a tool and you know there are great a lot of guys like Bill Gurley analysts and VCs have talked about this and you know listening to them this and this kind of rich history it it just makes me think in crypto we're we're a little bit naive

I think about open source software sometimes and we've overgeneralized and learned some of the wrong rules. here. And I one um one pretty persistent strategy that big corporates will use is they will leverage open- source to neutralize um the the a competing product or to you know to use you know investing speak like commoditize your compliment. So there was a great example when cloud was becoming a thing and you had AWS lead the way but then Google you know kind of fast followed and actually Microsoft with Azure and um there's obviously a hardware component towards building out these cloud data centers but there's a software component as well and soon everyone kind of became convinced right that there's not an

enormous amount of differentiation in between the underlying cloud computer the software I is a it could be a big lockin area and if you build your business on you know AWS how easy is it to port into another cloud provider? And Amazon anticipated frankly a lot of I mean I guess they didn't necessarily anticipate it because you already had Apple and that like um you know app store lock in but the thought process was okay if we think that we're heading towards a future where there's a low switching cost between different cloud providers. If we build an enormous amount of software like tools, libraries, whatever it is to lock it in and make it very difficult to switch.

That sounds really great. That's I could think I could juice my margins. I could do what Apple was doing and juice my hardware margins or whatever it is. Um and so Google rushed to market with Kubernetes. They didn't want Amazon to be able to limit the switching costs. And so they launched Kubernetes, this open source piece of software to neutralize that layer um and to make it um you know the switching costs very easy. So, I think that's just a fascinating example of how, you know, and and of course the messaging is Google loves developers and we're doing this as a gift and it's free, but it was just a genius business strategy. I think that's fascinating. Um, as an example here.

>> Yeah, exactly. And I think, you know, Facebook's doing kind of the same thing with Llama. Um, I haven't been as close to it. Um I wonder is it a little spicy take like is it is base a version of that and taking the EVM and that's obviously led to more sort of monetization for Coinbase taking that away from Ethereum even though Ethereum core devs are like roll up centric roadmap go do this um >> I think in hindsight that's like a maybe a bit of a stretch and a lot of nuance there but like yeah if you squint maybe fair I don't know if you guys agree.

Yeah, I think maybe you know it's been a great sort of historic view of open source and you know where winners have come out from open source versus closed source etc and business models there. Interestingly enough Miles you just mentioned like Llama from from Meta. Um without going too deep on the crypto and AI side uh yes on the open- source AI side what's happening today is Llama did or Meta did come out with that business strategy but ultimately actually very recently they decided to kind of dep prioritize that path because they realized that closed source models were just getting too far in front and even though they were trying to dig into the market share of um you know open AI etc it didn't make enough of an impact and

so it is interesting And I think we're seeing it in blockchains today as well. You know, we have these major layer ones like Ethereum and Salana etc. which actually do align a bunch of incentives and have a lot of people contributing etc. And Miles of course you're very very close to this at Delta with regards like how do you align incentives? How do you get developers on board etc. And I think Mike you mentioned it before with regards to like switching costs. And so again taking back to what you said as well Miles around like base. So you have these open layers in crypto which are like credibly neutral like Ethereum, Salana, etc. And I would say actually that base in some sense is actually taking the open- source stack of Ethereum, you know, especially the EVM and then monetizing that on top with like their existing distribution. And so

I think you see many many parallels actually in open source software that you guys just spoke about and crypto today. And I think we're maybe halfway through that transition in crypto right now with regards to actually monetizing some of the open source software that's been created in the past. But it's still an open debate I think right now. How can you create lock in? Because it goes against crypto's fundamental values. And I'm sure we'll touch on that in this episode with regards to the business model itself of trying to monetize open source software. And creating lock in goes against the freedom and permissionless of crypto as a technology. So yeah, it's a it's an interesting space in crypto because it's one of the only technologies ever in open source software which has made it possible to actually align incentives and create a business with. But then at

the same time, we're now seeing the reverse is true where you can also monetize it, centralize it, and go against crypto's ethos and still have a good business. >> Yep. 100%. >> And yeah, maybe to take like a little bit of a step back and then we'll come right back to that. save. Um I think like uh the open source communities like outside of crypto and open core has always been like very interested in crypto um and vice versa of course but I think they you know I remember listening to this uh awesome podcast with Nick Carter and Joseph Jax. Uh he's the founder of OSS Capital from like I don't know 2018 2019 um on Castle Islands pod.

Not to plug another pod on our pod but it's a good one. Um, and I think at the time we were looking at um some of the trade-offs of this open core strategy, right, which always ran into these like um tensions around what to try to monetize, what to, you know, keep open source and how to um, you know, relinquish governance but keep sufficient control, um, how to, I guess, not like lose that community that got you there in the first place. Um, and they looked at L1s in general, like because at the time it was mainly just L1's out there as potentially a breakthrough here, right? And Zave, I

think you just referenced this a little bit like for the first time, um, let's say Bitcoin and Ethereum was the representation of open- source software where monetization was baked into usage like um, you have first of all like blockchains introduced verifiability and like hard guarantees to the internet. Um and so you know unlike the internet anybody could if you ran the software and you connect to the network you can verify that everybody is playing by the rules right but to do that right it has to be open source in the first place right so all of L1's are open source because that's required to fulfill this

like promise right of adding guarantees to the internet but on top of that there's a single instance right of Bitcoin and Ethereum that is valuable it's one shared state network and to leverage the services provided by this network whether that's for payments or holding this like nons sovereign store of value you have to pay you have to pay the operators of the network um and I think with EIP559 >> with the burn associated with that right now that's actual direct value acrruel to token holders so you have a development company that puts out this blockchain they own a bunch of tokens that represent some you know stake in the chain either like direct control for

like POSOS networks um or you know a say over like who can run validators or something like that right and they if there's like monetization to the token or value acral to the token baked into usage of the blockchain uh we no longer have to build like a commercially viable separate company >> on top of the open- source software we can just hold these tokens and if it's a great product like all of the people who built it and hold those tokens get rich.

It's fantastic. So that that was like actually the primary breakthrough uh and why and why the space was like looking at crypto like oh my gosh maybe they they've actually got something really interesting here. Um, no, there's lots of nuances and like even within L1's where monetization is baked into usage like um we know that at the very least this is like for DOS protection so people can't just like spam the network. >> I was going to say the it's funny right >> well that and also the burn was originally introduced as a UX optimization. It was a uh you know the auction that people used to have to people forget before 1559 the auction mechanism of getting a transaction

placed on Ethereum. It was crazy. You had to estimate the gas. You had no idea what I mean. So it was really a spam prevention mechanism and it was making it was the the nuance there is it was a UX feature although you didn't have to burn you didn't have to burn the base fee. You could have given it to the next there were other schemes, right? But the burn was chosen uh to prevent off-chain validator uh collaboration and uh and it ended up looking a lot like value acral. I've flip-flopped on this a lot. I've been like h I think that's overhyped.

But now I I think it's very clean. And Miles, I just want to I I want to make sure that folks didn't miss that because this was the fundamental innovation of at least general purpose smart contract platforms like Ethereum. And Bitcoin does get an enormous amount of credit. It absolutely uh should. It solved huge computer science problems, the Byzantine general problems and proof of work. Those are massive innovations that led to Ethereum. But Ethereum itself deserves an enormous amount of credit for breaking this trade-off around monetization of open source software because Bit Bitcoin hasn't really solved that. It its design is, you know, things like fees and REV are naturally it is unoptimized for it is a big meme coin. I

think it's a very compelling meme coin um just like gold is, but there's no hard productive value. That's not to say it's not real value. It's just it's valuable in the way that precious metals are valuable. Um, Ethereum is very different and it broke this tradeoff of you have something that's open source and so you get free contribution and um and people want to use it and buy it because it's neutral. Uh, you you get that trade-off, but then you don't have the negative trade-offs of, you know, closed source of it being monetized and yada yada. You can have both through this this monetization scheme of you need to pay for usage in the native token. Um, for what it's worth, this is also why I'm not a big fan of people talking about Ethereum's monetary

premium like it doesn't need all these. You have this beautiful beautiful mechanism, right? You solve this unbelievable challenge. The the uh the ability to generate fees should be massive here um and super exciting and I don't understand why people continue to short change Ethereum on in that regard. But that is that is the fundamental trade-off that it broke. And I think that crypto has a really bad habit of taking like overgeneralizing solutions from the early success cases of Bitcoin and Ethereum into a bunch of places where it doesn't belong. And this is where I want to go into like L2s and infrastructure which I think just had the wrong mental model. And the my favorite one on Bitcoin is the 21

million hard cap people have applied to applications. It's like you don't need that. Um, we don't need to reinvent equity financing. It's much better to not have a hard cap. But I think that people took this this idea the the breaking of the trade-off and paying in the native token as a way of monetizing open source software and they overgeneralized that to applications and additional layers of middleware and infrastructure where it didn't belong. This is where you get this whole idea of a utility token which is a broken idea and and doesn't really make any sense. Uh, but Yeah, to be fair, like they broke the Owens broke the trade-off of like uh uh the value acal trade-off associated with open source software. I don't think they

broke the governance trade-offs. Um cuz I the other one I mentioned is like the more you lean into like open- source, you know, ethos and decentralizing control uh and like of the open source project itself, the harder it is to be competitive, right? So, I would argue that like Bitcoin and ETH have been, you know, they had like ideological roots and they've stuck to those. Um, and it's been challenging to stay competitive with, um, I'd say like new open-source blockchain projects. Um, say like Hyperlid and Salon are probably the best examples of like breaking both trade-offs. um they've managed to find a

community of users that like are okay with them retaining enough control to stay competitive and they've implemented you know the sort of value acrruel like baked into usage but um then yeah outside of that like there's a ton of ton of you know uh issues I I think it basically gets really hazy as to like whether or not like we um break these trade-offs outside of L1's. I would argue no, but I think everything outside of L1 still should be most of everything outside of L1's in crypto if they're part of the value prop of providing verifiable services like they should be open source. It's just monetization is not as naturally baked into them. Um,

>> remember like Zerox, like one of the oldest AMMs or Dexes in the space originally had people paying fees, you know, in that token and everybody's realized that's a really really bad idea. Um, or it's just like a bad product, right? Um, but yeah, Z, I know you want to jump in. So, please do. Yeah, I I think maybe just to paint a picture here, one thing that's interesting which has been happening in crypto the last year or two especially or few years I should say is in the open- source landscape Miles you mentioned Salana and Hyperlquid as well uh coming up as you know major open source layers in crypto which is definitely true however at the same time Ethereum and actually Vitalic came out with a tweet about this you know very recently I think it was this week around

the sort of northstar for Ethereum as being this incredibly neutral layer and how important that is for Ethereum and he coped some criticism for that. Salana was built as an open- source software stack that was a bit more top down. So yes, it's open source. However, there are like little colors of closed source in Salana, especially at the application layer. So when we used to work with Salana apps very early on when we were building light back in the day at Chorus one, a big criticism of the Salana ecosystem was actually how they were closed sourcing the apps. But it's never been significant enough for people to care and stop building on Salana. And so that goes to show that base layer being open source is actually theoretically more important at least the market decides that right now than the applications themselves. I also want to

like point out Hyperlquid here cuz yes again it is another successful open- source layer. However, a lot of the binaries etc that they were giving to validators were not open source. And so there's a lot of hyperlquid actually that is not open source. And so I just want to like actually point out this trend which is I do think that you are seeing this trend in crypto right now at the open source layer at you know at the credibly neutral layer of actually making certain parts of that more closed source over time to capture more and more value. And I would make the argument that Ethereum actually is still potentially the only one with the north star to be a purely open- source layer.

And again, we've spoken about this on the bell curve in the last few weeks, but I do think that's interesting to point out that you're now seeing a divergence between new layers trying to top down close to a certain pass and capture value versus like the Ethereum style of incredibly neutral north star. >> There was a debate that went late into the night of block works last night. Uh we launched this data website and there was a debate around how should you treat the rev uh fee stream of something like Ethereum compared to the revenue of something like Hyperlid. um from a valuation perspective like one do you put them in the same bucket um like a hyperlquid and ethereum like the rev gets equally paid out to token holders and it's generated in this very different kind of holistic network rake

sort of way versus the revenue stream that comes from hyperlquid is derived from the it's it's almost like having a very successful owned app on top of the ecosystem and so should you treat those the same anyway there's there's like an >> we're not going to go down this rabbit hole there's a deep rabbit hole there to go down. That's >> Can I add Can I add one little like uh comment on that? >> Yeah. >> A little bit closer to this conversation. I'm um I've seen like L1's be really successful from a monetization standpoint while being open source uh when they have MEV and they host execution directly or they host an app directly, right? Um the you and I did a bunch of podcasts on

like Celestia back in the day like modular blockchains and there was this idea of like um and maybe Ethl1 like uh for rollup fees is in the same boat of like Costco right where they're you know kind of like um not able to monetize directly. Yeah. Great business. >> So like yeah no that's a great like don't worry guys like we'll monetize like Costco right? We're not like because I'm building a similar like Delta is similar like that. Like we do we very purposely offer apps to control their own execution environments and you're just paying for like settlement right which is b basically like aggregated you know state updates and there's no easy way to price that because you have no you can't just be like yeah 1% of your revenue should go

to the base layer because it's that's not the way like these these modular blockchains work right so maybe just as a fun like because I love nerding out on like blockchain financials I think that if you host execution directly And there's ME across a lot of apps. That's been proven to be like very monetizable. If you choose to host a blockchain with a single app, like that's been clearly monetizable. Um, and the jury is still out on I think like um any L1s that don't host like execution directly. I mean, look at Bitcoin. It's the biggest one. There's no execution really.

There's nothing happening over there. So, I don't know. Sorry for the tangent, but yeah, >> we we have our whole story or our whole episode to do on buybacks. I have a lot of thoughts about this, but um let's I I want to get back to layer twos here specifically and I want to talk about this tra. So, okay, let's go back to our point in the narrative story here. We are very excited because historically there's been this trade-off of you get these wonderful things with open source, but you can't monetize or you get closed source and it's very difficult to build community. Amazing. L1's introduced this great trade-off where you pay for usage and you seemingly get in some cases maybe governance excluding air miles the best of both worlds. So we start to over but maybe because governance is so slow we can't move fast it's oified we start

to get and fees are high on something like Ethereum we start to get these pro uh let's call them vendors uh because I think it's going to dump down nicely these vendors that build on top of Ethereum and they solve core problems with building on Ethereum. So that's everything from it's really cost um confirmation time um but there are other uh actually it creates issues around interoperability as well but uh you really it ended up being around speed and cost I think um and so a lot of these L2 I don't know if people remember L2s in the beginning but a lot of people were questioning why do we need a token for this um and there were questions around where was the revenue stream and

the lock in the the the token sync going to come from and where are we going to pay for sequencer fees in the native token and and things like this and I think that a lot of the business model from business model for Ethereum people generalized to L2s and at the same time you also had uh this was when the modular meme was taking off and you had these other services that were popping up to solve the problems with L2s so okay L2s are not interoperable with one another so you get shared sequencers And by the way, it's actually not even super easy, even if the L2 is fast and cheap to build there. And so you might need these other services that make it

super easy, like a RAS provider. Um, and so you had all of these different infrastructures pop up to provide this patchwork and solve this problem. And many of them were operating with a lot of either explicitly or implicitly the model that Ethereum had set. So, we don't really need to monetize. We're going to have a token that we're going to stick in here and the usage will pay for all of this. And we are now in it's January of 2026 and I'm curious, you know, if we were to do a retrospective on L2s and how their business models have evolved. There's a lot of really interesting stuff, by the way, we could talk about we have on this podcast in between the different model of an optimism versus arbitrum, um, which had

very different approaches. But, you know, what's what's our retrospective on L2s? Um, and maybe we could start to knit this into, you know, some of the analoges of like the Red Hats of the world, right? And some of the way that other commercial open-source software companies are monetized. >> Yeah. I'll leave it to you guys for the analoges, I think, but I can give my view on, I guess, L2s and where they're at right now and where they've come from. So I think that when L2 started going live, you know, the reason why they went live in the first place was because Ethereum wasn't scaling to be honest. And so Ethereum deliberately tried pushing this narrative of like we want really strong builders to develop layer 2s to get people off the base

layer to make sure that the base layer isn't too expensive for like the operations that matter the most. I would say like since then especially in the last 12 months it's very clear that Ethereum is trying to move towards the base layer again as they continue scaling ZKVM etc etc and so even when they push they they tried pushing activity to the layer twos you're seeing now that like a lot of activity just stayed on the layer 1 to be completely honest at the same time that Ethereum drove that layer 2 narrative Salana popped up and they have this whole layer 1 narrative of like actually keep everything on the base chain and we've seen how successful that is so in some sense like yes Ethereum is similar to the analogies we were drawing on before of like open source software people

building tools on top of that different business models I actually think where it gets quite interesting is like we have seen some layer twos with some success but there's a power law and like that was pretty clear from the start that was always going to happen and that's why so much venture capital funding went into this because everyone was seeking that power law in the layer 2 landscape of like there will be one or two layer 2s which have enough of a user base and liquidity and state to actually make it and continue being significant as we keep going forward in crypto. those two layer twos right now really I think I mean we have different opinions on these two layer twos but one is arbitum and and one is bass and I would argue that arbitum is more of a kind of defi chain base we've described before

in the past that it's kind of working out what it is I would argue that bass right now is kind of more consumer appy moving towards crypto AI some payments etc and then you have ethereum the base layer the layer one I think Ethereum is like very very clearly right now like a defi chain and I think all of us can probably agree on that as well but then you kind have those bases covered in like what a blockchain actually needs. I'm not sure Ethereum base layer will ever be like the layer that consumer apps are built on top of versus like a Salana for example, but I do think that having that layer 2 there is very important like a base that then rolls up to Ethereum base layer. I think where it gets interesting is like you know we didn't talk about this before but

because Ethereum is still so you know decentralized and we're still working out what the value of that actually is. I do still continue to make the argument that Ethereum does have some monetary aspect to it and I do think it's one of the most credible assets we can talk about right now outside Bitcoin that does have a monetary aspect and that's hence why I do believe why it does actually have some monetary premium when you look at the other kinds of assets like hyperlquid etc. It's not as open and open source and transparent and decentralized etc etc and therefore a lot of the reb like you were saying before Mike is actually valued in a different way. It's very much like a normal business would be. Layer 2 is the same where you have one node taking all the transactions all the data you know

making sure it's verifiable rolling that up to Ethereum but then their business model was trying to capture and take as much from Ethereum as possible which makes sense. So you have this kind of like I think we're in this place right now where Ethereum and Bass would actually end up competing with each other as layer ones and layer twos. And so this is where it's going to get very interesting where we're very early in what this looks like. But I could see a world where you know base takes so much power from Ethereum where they actually end up creating their own layer. And then Mike you've said in the past before like can you actually do that or not? I think it's TBD. I remember back in the day in the layer 2 landscape you had like starknet etc try to build layer ones almost. So they tried exploring

what a layer 1 would look like. A lot of that was using the Cosmos SDK which of course all of us here are familiar with. They kind of have moved away from that ultimately and so there was a lot of talk back in the day of decentralizing layer twos and that looks a lot more like a layer one when you have a decentralized layer two but right now that hasn't happened. It's two or three main layer twos. I think they will continue winning. I also think Ethereum will continue trying to take what they lost from these layer twos. And then I actually do think that there will be some competition between especially Bass and Ethereum and there's some tension there. But yeah, curious to hear you guys see it.

>> I well I I have I have I think I would just introduce I think the mental model that was not necessarily correct and I think the the jury is still out on who is going to win this. I would I I would actually add to this pile. Um, optimism I think is still very much in the running here as base is on optimism on the OP stack currently and they've been really successful at onboarding, you know, new chains like Inc. uh, from Kraken and things like that. So, you know, there's they're very much still in the running. The other, you know, interesting area where you're starting to see convergence and Miles got to give you credit for this, you were talking about this two or three years ago, but uh, we just saw ZK Sync uh, for their uh, you know, um, elastic chains. they

they launched their own RAS like integrated service and so people are coming I think people are starting to arrive at this solution and again I the analogy here I I very much view these businesses like Red Hat where ultimately what people want what your buyer wants is because of this open-source also like real network of liquidity users people want Ethereum but they don't want the downsides of Ethereum namely the cost you know Ethereum doesn't go down, but gas fees certainly do spike to a degree that it would disrupt your business operations and that's not going to fly.

So, I mean, it's just super interesting to hear how, you know, Red Hat packaged Linux, right? Like the CIO wants Linux, but they don't want any of the risks of uh, you know, downtimes. Um, they need SLAs's, right? They need to know who they're going to call at 2:30 or 3 a.m. You can't call anyone from Ethereum at 2:30 a.m. You might be able to call someone from an ecosystem like Mega Eventually. I I don't know. Um and so I think that those exact same So I I think that the the mistake that a lot of um these L2s made was I think they went a little bit too hard on building their own ecosystem when really what they were was they should be leveraging and they kind of did this naturally, right? like

they they all like we are Ethereum but they should focus on repackaging Ethereum to make it attractive as opposed to I don't know the the the jury's out on whether you're able to build their own ecosystems and I think that'll actually be the challenge for B safe to flip it around is like I'm actually it could go really either way right and you've seen different rollup frameworks take opposite strategies here of I'm you know arbitrum one has been unbelievably successful like they I think are kind of the only L2 that has this really like deep and liquid and used uh DeFi ecosystem which is super valuable and it feels like other L2s have taken the strategy like optimism abandoned OP main chain um and they said no one's going to use this everyone's

going to want their own chain and so frankly the jury is very much still out on there's probably a path for multiple chains to win here I'm sure there will be but yeah >> um I think the model is less ecosystem more Red Hat >> right I think it's like um they all came from the same starting point. You develop a rollup framework, right? Um and then you have two paths. You can build a first party rollup, right? That's OP1 uh or OP mainet ARB1. Uh and that's easy to monetize. Um it's just like an L1 except you don't have to share it with validators, right? Usage fees accur. But the more you prioritize your first party rollup, the less that

you can really optimize for wide adoption of your rollup framework, right? Because if you are like um pouring all of the resources and growth marketing into your first party one, the less likely a really high quality, you know, uh company is to adopt your framework because they don't they know that you're not going to be you're going to be competing with them instead of prioritizing them, right? Um but it is much more simple to monetize. Um, so I think you know you could argue that like ARB one is basically focused on that even though they haven't really but that's the way it's played out. Um then optimism I think you know to your point looked at this thing and said well there's no way that OP mainet is ever going to compete with Ethereum right um

we're not going to try to um we're going to try to have like as many people adopt our rollup framework as possible all for like different use cases and different companies different flavors right and then we're going to connect them all um and the I think dream formula here which is finally being done by ZK DK sync is that um you know it's damn hard to try to monetize a rollup framework without like a RAS solution. Um it's just you're not giving enough right like with a RAS solution you are hosting the rollup for them. So it's no DevOps headaches.

You're keeping up upgrades in line with like the core protocol. You are building indexers. you're building um plugins for every single critical third party like you know service that they it that they would need anyways and instead of you're taking that off the plate like the discovery work of going and finding those it's more like if you go to conduit's website it's like boop boop boop all right got my roll up here um and so I think the big mistake was that the equity company should have launched like the dev code that built the open source rollup framework should have built their own RAS solution they can monetize the equity company from that RAS solution. Sure, at some point down the road, you want to start doing

buybacks of like a governance token from it. Great. Um, what I think what ZKY Sync's doing, which is super elegant, is that uh token value acup uh between the different frameworks and it's baked into settlement to like a common, you know, L2, right? And they're monetizing like the hosted services with the RAS business and, you know, maybe that ends up buying back tokens. Maybe it just is used to like fund growth and development of the Devco, right? Without having to ask the DAO or the token holders for money, which is really really elegant. And so this to me was like um why isn't anybody doing this?

Why why is OP like letting everybody eat their lunch back in the day? Um, and I think you know Zab to your point like the funny thing is if you really want to optimize for that first party roll up and make it as big as Ethereum like the logical thing is to decentralize that roll up right so that is censorship resistance even like in a short-term CR standpoint as as a blockchain and nobody's done that yet. Um and so what has also been interesting is like you're right Mike like OP basically had they said oh Bass would might adopt our framework like yep base we are going to dep prioritize all activity on OP mainet this is going to like we're going to push that to you um and that was probably part of like the un unspoken

agreement doesn't seem like Robin Hood's you know arbitum framework has really siphoned off that much activity at all I don't even know what the status of it is frankly um from ARB one But yeah, I think like uh I think we'll see like the the logical like um nice balance of token, you know, usage uh value acrruel, off-chain revenue that's geared towards like, you know, broadening adoption with with ZKYNC's um balance. And yeah, I think that's one of the things I'm paying attention to. Um it's taken a long time. Maybe just to chip in briefly, I think where you don't want to be right now in the open source stack is like in the middle. So like I would make

the argument that deploying rollups on behalf of people and finding other people with large user bases to then support, you don't have any economic power over these teams because they own the end user. And then on the flip side, that settlement layer of Ethereum, you also don't own that because you know, you're never going to out compete Ethereum like in the EVM ecosystem open source right now. So I do think there's some tension again here around like yes there are now roll up as a service layers but do they actually acrewue any value and how competitive this whole stack is where the next person comes out with the next framework and I also just wanted to briefly touch on this as well you don't want to be the cosmos SDK of blockchains because I think like where a lot of

>> plus these rollup frameworks are kind of trending towards because cosmos SDK okay what was it it was an open source stack and it was beautiful and you had interoperability, but it sort of failed or did fail to capture value across the entire Cosmos SDK stack. And I think a lot of these RAS frameworks could suffer a similar fate ultimately because unless they actually own this end user and of course the Cosmos SDK, you could own the end user at the same time as owning the base layer. That was the beauty of Cosmos SDK and that's what Hyperlid is actually building on top of right now. to some extent it's like hyperlid because it had the state it had the liquidity it had the users it had the adoption it had the distri it it had everything the underlying stack the people talk about right now in crypto

Cosmos SDK was the best stack to build hyperlquid with like no one really talks about that so I do think the risk on some of these like more RAS providers like optimism etc is like that's great there could be the underlying stack and they might try and make that their business model but how much are people willing to actually pay for that versus like just building the stack on top of them and trying to take as much ownership as possible So, it's an interesting way of viewing. >> I think I think it all comes down to like is there one instance of this thing that everybody uses? If so, then it's actually very easy to monetize through usage, right? I could argue the same thing for like a and unis swap like like you could have forked those a long time ago, but there's one instance that had

the liquidity, right? So, people will pay to use it. Cosmos SDK there was not like um it was an internet of blockchains by default. And then it's like um you're kind of looking for ways to back into monetization from that. And it's been really hard like um you know I think like the IBC routing hub was one idea and usually it comes down to like interop between these different instances that people try to like land on but hasn't been simple. Um, and then I would also argue like a lot of middleware has taken this like trying to be like POS approach as well in order to like have some sort of monetization, you know, that's baked into usage. Um as I can tell you like as a like we're

building um a lot of like Delta is built uh heavily with SP1 >> and you know you can run those proofs locally um as a domain on Delta or you could route proofs to the succinct network. Um and as like you know one of the builders of Delta I was pissed when I found out that like you had to use prove tokens to use the succinct network. I was like nobody like who's just we just want proofs like this does this can just be like a SAS solution right um and of course like they have other sort of monetization like yeah you know they have other sort of pricing models now and I you know got walked back off the cliff but it that that is kind of like the uh the skuorphic like

we have this we want to token we want to we we see this like elegant usage thing and then like that just turns out to kind of degrade the product if you push it too hard. Um, >> yeah. >> So, >> one thing that I and I want to like transition us here to AI and what's changing here. So, like if I just had to summarize this, right? Like the throughine of this episode is software, how it's developed, the different business models that have arisen around it and what changes crypto and eventually AI are bringing to open source software specifically. And I I think hopefully some of this has been enlightening in terms of Miles is your observation and I completely agree. I think the L1 is L1's are unique in the sense that it they broke this trade-off.

I think we applied that model of L1's to a bunch of different layers of the stack especially around infrastructure and it hasn't worked for a number of reasons. That's far from saying that these companies are not going to work. I I think that they will and they'll be very successful, but they need to there is going to be consolidation would be a prediction and you you might actually see a change in business model. It's possible that um maybe they could optimize for something that looks more like a recurring revenue sort of SAS stream. The only thing that gives me some amount of pause in hearing you say that is at the same time that we're talking about enterprise and B2B SAS that whole industry is getting that whole model is potentially at risk from AI. And maybe whoever wants to take this

cuz I feel like I've been monologuing about this episode, but I mean >> Z lead us to the promised land. Give us how how should people think about AI and these LLMs impacting B2B SAS as a first of all why is it so disruptive? Why is it potentially disruptive? Why are people concerned about this? And then how is this impacting business models? Yeah, I think in general of course AI is impacting many many industries in in different ways and you know to make it a bit more relevant to crypto and AI I guess which is sort of the crux of this podcast you know there are right now with open AI you have chatgbt etc you have these massive models that pretty

much all of us use today Google gemini etc um and these models are closed source and actually there are many parallels you can draw on AI to kind of like finance actually of why crypto was created in the first place. And really ultimately there's one word for this which is trust. And so you know we lost trust in the financial system you know back in ' 08 global financial crisis. And that was when Bitcoin was born you know an open-source stack for finance and and money and then Ethereum built on top of that and we've made many many uh improvements since that that date. So I think with AI we're going through a similar thing right now which is actually that a lot of power is in the hands of the very few like we had with the financial system back in 2008. Maybe

there are still some massive power sort of pockets in in the world still which they ultimately could still go down and make a big impact but we have this alternative financial system which like we can still use regardless of where the power falls across the world. with AI like I think it's a very important initiative to continue looking at open source through blockchain because we have this technology stack that is an alternative in case it's like this hedge against AI in the real world potentially creating something so massive that it's like actually impacts all of us and that's actually a non-zero% chance of happening and so what's actually important right now is that open source tries to keep up with closed source improvements in AI and you're seeing

there's actually a curve of right now which we can maybe put in the footnotes of this podcast where open source AI performance is actually catching up to closed source AI over time as closed source now is a bit more incremental when it comes to improvements whereas open source AI for example most of us heard 12 months ago when Deep Seek came out with you know open source weights and open source uh models that that was like pretty much almost on par with like some of the closed source models at the time. So there's a lot of there's a gradient and I'm sure we'll touch on this in the like the kind of crypto AI episodes we we go to, but this gradient for open source AI even meta which we mentioned earlier this episode they were somewhat open source but actually they were open weights and so it's still very

hard to decipher exactly like what was going on in the model, what data were they like using to train their models etc. And so even Meta when they were saying they're open source is terminology but it's a marketing thing of like saying they're open source. So, crypto really brings this pure open- source like version of AI to the world. And I would say that there was a lot of hype in open source AI with crypto incentives. And a lot of this draws on what we've said in this entire podcast around like using incentives, using a network for usage, aligning incentives with contributors, having anyone contribute to an open source stack. And actually in that way theoretically this market can actually value what these AI models are actually generating as

outputs which we don't have anywhere else actually in the world today. And so more like recently I've actually been spending a bit more time in the Bit Tensor ecosystem. I think this came out all of a sudden and there was a lot of AI hype and then Bit Tensor kind of popped up as this ecosystem. And then what's very interesting I find about Bit Tensor is that you have kind of people really solely focused on this ecosystem. So it's a large enough ecosystem in crypto for it to matter enough for people to build funds around build ecosystems around you know DCG is is very involved in this ecosystem for example and what Bitensor is as like one way of open source AI is really contributors giving their models and trying to respond to you know user inputs in the best way possible and actually competing against all the other

miners let's say in this network then you have another layer on top which are validators which are ranking these responses and then the best responses are ranked using this algorithm And that is a consensus and then that is a free market way of responding to user inputs using the best uh outputs of the network contributed by this miners open source. Theoretically they're always going to be incentivized to contribute their models because there's tower tokens being you know there are tower missions which is the token of bit tensor going to the best models and so you create this free market for AI actually which is very early still. I think there's like maybe 100 plus subnets at the moment and each subnet is specialized in something different in AI but in general it's one way of looking at how we can think about AI as a free market uh which is still

early and the question is like are the responses of the closed or the open source models actually good enough versus the closed source models because most of us just use chatgbt or something right now which is just handles a lot of this for us but bit tensor and open source AI is getting to the point now where like open source performance a bit like money and finance has in blockchains, you know, across the last 10 years is catching up to closed source. And so I do think there's a lot of things to be said around open sourcing AI, using incentives, aligning contributors to the ecosystem, having code. A big thing with open source AI as well is actually just transparency and auditability, which you didn't have in finance for a long time. That's actually what sort of brought the whole ecosystem down in 2008. So if you can bring

explanability transparency auditability to AI, which actually isn't there today, I think this would be huge in like the long long term. The question is like where is the value today? And I think we'll explore this in the crypto sort of like themes and topics we touch on in the next few weeks. But there are certain subnets doing actually quite well in Bit Tensor. There are things outside of Bitensor right now happening in open source AI which don't touch crypto as well like for example what Meta has done which we touched upon and some other things like Deepseek. But I do see a convergence of open source AI like deepseek with blockchains for example you getting these models on chain allowing people to contribute data for example which of course is a bottleneck in close source AI and we keep saying that this is a problem maybe

in the past it hasn't been a problem because there's been new techniques developed by AI companies but at some point there will be a lot of value in contributing certain amounts of data and getting sort of paid for that data and that's a very cryptoreated concept there's another point of like models themselves how they're commoditizing anyway way in in the closed source world. And so if you have a commodity as a model, bringing that onchain for everyone to use and then building on top of that to capture that value, I think is a very very you know possible way that that that apps could monetize in future. They won't need to use open AI APIs. They can just use actual open- source APIs via blockchains for much cheaper. And so you can see ways that the business model can change in AI. is still like very very early on in its

sort of lifetime, but I do see a viable path actually as open source catches up to closed source in AI for it to become this massive thing in crypto as one of blockchain's actual biggest use cases. But again, it's very very early. Um, and again it's like I maybe would make the prediction which might be a bold prediction that an open- source AI company like a DeepSig or like something else will try and monetize via blockchain at some point in time because it just makes sense like we've talked about in this podcast about using a blockchain as a way to monetize open source software. I wouldn't be surprised if like actually one of these models came out eventually when blockchain technology is actually good enough to for people to use to monetize in that way. So it'll be interesting to see how

like time frame that happens. >> Yeah, because that would be more like a like a depend play or just like a livveness and like verifiability play. Like it it's not actually like a blockchain in the classical sense, but like all the validators are running the model twothirds sign off on what they return to you. Um and yeah, if that can be competitive with just, you know, running a single instance of it, we'll see. But um >> it could also be like a roll up realistically like you could just have one or two nodes like running the model and as long as it's kind of verifiable when the data is available for example and you can check that on chain and then maybe reward some kinds of contributors.

I've actually spoken to a lot of teams working on like contributing data to different models and it's really hard to do at scale. So the question is like can they do that? Who will do that? And how much value is there in actually having an open source model which is run with similar performance to closed source which people continue to contribute data to and like can that be a viable approach to AI outside of what we have today. against >> super super interesting and that's why I'm not sure how many episodes we're going to do on AI but like we need need to have more than one because like it is uh there's intersections you know with crypto or blockchain at like every layer of the stack from like model development to like who's actually running the model like making the models like pushing the

models to be somewhat deterministic like that's kind of what I's doing um and then yeah we can talk about what what we're doing at Delta and like a few others are in terms of using like blockchains to anchor guard rails, right? Um, so yeah, >> you know what I would you know what I would also say is I think there are three areas that I would point to, Miles, to go to your last point there. I think eventually doing things like spending money on chain, there are fewer hoops that you need to jump through, like banking APIs that you need to call. Money moves, value moves easier on chain than it does on the Trady patchwork L1.

And I think that will eventually be a big advantage. we talked about. I don't think it's maybe as soon as you guys do, but I do see a huge uh advantage there. The other things are a little bit murkier, but you know, we've talked about crypto unintentionally sometimes foreshadowing the future. And you know, Miles, we were talking a little bit before we got on that how do you build a moat even in closed source software like SAS today if you know soon AI is going to be good that you can essentially fork closed source code. And I think that people will learn from what's happened in crypto. And we could do a whole episode on forking. And I think the the business model of how you build up and retain value and build a moat. Um, you know, I

think traditional software companies should look at crypto who've been compete who has been competing in this environment for a very very long time. And I think that there's a real connection there. Um, and then lastly, there is a a model in the pricing model in SAS is changing. Um, and it's going from, you know, monthly or quarterly or annual, uh, you know, flat recurring fees into something that's more usagebased. And it's interesting like when we go back to L1's and their model of essentially usage based fees, that looks a lot like where we might be headed with these LLMs and AI. I mean, it is interesting, right? that I don't know if you ever if any of you read and we should probably jump off of this cuz it was so good the rivet piece on tokenization connecting the idea of

tokens that are used in LLMs with tokens on chain. It is fascinating. Uh but I think that's the ultimate connection there is you can envision a world very easily where everything is this digital representation and you know you're feeding things into an LLM to help you get an idea and actually the usagebased model maybe with like a a margin on top makes more sense than a SAS model. So the world is the world is changing really quickly and it is super interesting and I I I think you know we wanted to introduce we we had this idea for doing an open source software you know history retrospective and how it fits into the evolution of crypto and AI. But I think that's a really good lead into the next couple episodes that we'll do on AI and the intersection of AI and crypto. But I know guys we've

already gone a little bit over but this this was a really fun one and um I'm looking forward to running this back uh next week with AI AI crypto. >> Let's do it. Thanks for having us, Mike. It was a great episode. >> That was a blast. >> That was great. Bye, guys. >> Cheers.