Crypto Experts Explain Stablecoins & the Future Financial System w/ Ali Yahya & Arianna Simpson

a16z Crypto General Partners Ali Yahya, Arianna Simpson, and Erik Torenberg break down what’s actually working in crypto today - starting with the rise of stablecoins as a real-world payments layer. They discuss how stablecoins are being adopted by companies like Stripe and SpaceX, why regulatory shifts are opening new doors for crypto startups, and how AI and crypto are beginning to intersect. They also cover: - The future of decentralized social networks - Where Ethereum, Solana, and others stand today - Misconceptions still holding the space back - A grounded conversation on what’s real, what’s hype, and where crypto’s finally finding traction. Timecodes: 00:00 Introduction 00:45 Bitcoin's Original Vision and Evolution 01:47 Stablecoins: The Game Changer 03:02 Current Use Cases of Stablecoins 05:22 Challenges in the Financial System 07:08 The Future of Stablecoins and Financial Systems 09:02 The Role of Big Companies in Crypto 18:21 Decentralized Social Networks and Consumer Preferences 23:20 The Intersection of AI and Crypto 31:32 Misconceptions About Crypto 35:54 Smart Contract Platform Wars 39:33 Policy Changes and Future Opportunities Resources: Find Ali on X: https://x.com/alive_eth Find Arianna on X: https://x.com/AriannaSimpson Stay Updated: Let us know what you think: https://ratethispodcast.com/a16z Find a16z on Twitter: https://twitter.com/a16z Find a16z on LinkedIn: https://www.linkedin.com/company/a16z Subscribe on your favorite podcast app: https://a16z.simplecast.com/ Follow our host: https://x.com/eriktorenberg Please note that the content here is for informational purposes only; should NOT be taken as legal, business, tax, or investment advice or be used to evaluate any investment or security; and is not directed at any investors or potential investors in any a16z fund. a16z and its affiliates may maintain investments in the companies discussed. For more details please see a16z.com/disclosures.

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Ali Yahya?

  Crypto can help decentralize the power structures that are emerging in AI.

• 0:05Speaker 01

  Chris always talks about, do you want to be the indie band? Or do you want to play like the Super Bowl or the mega stadium? And I think like stablecoins really have the ability to appeal to like a much broader audience.

• 0:16Ali Yahya?

  Stablecoins are beginning to really gain traction. So there's something like 16 trillion dollars in volume on stablecoins per year.

• 0:23Speaker 01

  I actually think it's a great time for folks to be building token networks.

• 0:28Ali Yahya?

  Crypto is like a fundamentally radical set of technologies that is very, very hard for incumbent players to adopt and run with, precisely because it is so fundamentally disruptive to the way that they do things.

• 0:41Speaker 02

  What's actually working in crypto right now? For a long time, this space has been defined by bold visions and has elicited strong skepticism. So today, we're getting clear on what's real, what's being used at scale, and what's coming next. Joining me are two of my fellow general partners here at A16Z. Ali Yahya, who leads investment across crypto infrastructure and developer tools, and Arianna Simpson, who focuses on early-stage crypto networks and founders building at the edge. We get into why stablecoins may finally be crypto's breakout product, how AI agents are creating new demand for crypto rails, and what's changing in policy that could unlock the next wave of token networks. We also talk about the enduring vision for decentralized social,

• 1:22Speaker 02

  the evolving smart contract landscape, and why Ethereum is still widely misunderstood. Let's get into it.

• 1:32Speaker 02

  As a reminder, the content here is for informational purposes only. Should not be taken as legal, business, tax, or investment advice, or be used to evaluate any investment or security, and is not directed at any investors or potential investors in any A16Z fund. Please note that A16Z and its affiliates may also maintain investments in the companies discussed in this podcast. For more details, including a link to our investments, please see a16z.com/disclosures.

• 2:02Erik Torenberg?

  So I'm excited to do a deep dive with you on where we're at today in the space. So crypto is a space where people have long been excited about the vision and the potential. And people have long also been skeptical about where the use case is, what's happening, what's actually working. So here we are in May 2025. Why don't you give some context on what's actually worked so far, what's working right now?

• 2:24Ali Yahya?

  It's quite interesting because if you go back all the way to 2009, when the original Bitcoin white paper was published, one of the first few lines of the paper describes Bitcoin as a peer-to-peer electronic payment system, which was kind of the original vision behind what a blockchain could do. And it's really taken us like 15, 16 years to get to a point where the technology is mature enough to actually make that a reality. And this is now manifesting with stable coins. So some of the big issues that Bitcoin had that made it impossible for Bitcoin to become that peer-to-peer electronic payment system is that, one, it was extremely inefficient and very slow, and it still is. And therefore, it's become more of a store of value type of system as opposed to stable coins.

• 3:06Ali Yahya?

  And two, Bitcoin is not a stable unit of account. And so it's very hard to use it for payments. And so since then, one of the big things that has happened is that the infrastructure has matured tremendously to the point at which now we are at a level where a transaction of any amount of money can be done for less than a penny in cost. And it can be done in under a second, roughly, like those numbers are approximate, which finally makes something like a peer-to-peer transaction of a few dollars viable on the blockchain. And that combined with the regulatory clarity that we're having now as of the new administration makes stable coins something that are really beginning to happen. So that's maybe like the biggest thing that's going on in the crypto world at the moment is that stable coins are beginning to really gain traction.

• 3:51Ali Yahya?

  So there's something like $16 trillion in volume on stable coins per year. And there are many traditional financial institutions that are beginning to use stable coins to rip out a lot of the backend of their financial systems. And these are like fintech companies, think Stripe, think Revolut, think Robinhood. Some of the companies in the traditional financial system that rely heavily on the trad financial system are now realizing that stable coins are a much better way to do things. So that's kind of like the biggest thing that's going on. And we believe that will likely lead to this. Cascading trend of adoption because once stablecoins become a more kind of mainstay of the way that the financial system works, that opens the door for a lot of the other more advanced and futuristic ideas that crypto has introduced like DeFi

• 4:36Ali Yahya?

  to begin to also gain adoption. And I think that as a result will lead to all of the other things that we believe crypto can offer to really start to ramp up.

• 4:45Speaker 01

  One interesting thing, though, is that stablecoins, at least us who are in the industry full time, have been thinking about for like years and years. Because I remember talking about them in 2017, 2018. And I think there was always a narrative about them being useful for remittances or in countries that have had hyperinflation. And for those countries, Bitcoin is a better store of value than their native currencies because sometimes it goes up, unlike those which only go down. But it's not ideal because, again, as Ali mentioned, it's not a stable unit of account. And so it's interesting to see that even though this has been talked about for years, now it's really having its moment. And I think to Ali's point, a big part of why that's happening is because the infrastructure has evolved to a point where you can now efficiently move money and you're not having to spend a huge amount of money to move the money,

• 5:30Speaker 01

  among other things. So I think that's why we're starting to really see it shine now. I would also add that it's intersecting in interesting ways. We're still super early in this, but there's obviously a lot of talk about AI and agents. And if you want to dispatch your agent to go transact on your behalf, you can't really give them your bank account or your credit card, but instead you can give them your crypto wallet. And so this interplay of agents buying or spending money on behalf of their users with stablecoins is a really interesting theme that we're starting to explore.

• 6:01Ali Yahya?

  Actually, to that point, it's kind of ridiculous to think about the way that the financial system works today, where even a normal local, like domestic financial transaction, where you go to a coffee shop and you buy a coffee with a credit card, that transaction involves like the point of sale, the payment processor, the issuing bank, the acquiring bank, the credit card network. And each of these intermediaries takes some cut, some fee on the transaction to add up to something like multiple percentage points on the transaction. And that is the case in A domestic transaction and if the transaction happens to be international, then that entire stack of participants and intermediaries gets duplicated and mirrored on the other side to the point that any kind of financial transaction across borders is insane in terms of its inefficiency.

• 6:45Ali Yahya?

  It can take up to like three to seven days to move money from one country to another and it can cost like up to 10% of the transaction to do it. So when you have a technology that can now again move an arbitrary amount of money from anywhere in the world to any other place in the world for under a penny and under a second, that truly is very transformative and it'll be very disruptive to the way that the financial system works. So to Oriana's point about AI agents, it's kind of inconceivable that an AI agent that a human may want to participate in the financial system would have to go through all of that inefficiency and deal with all of these arcane human intermediaries, some of which are not even really automated. That's inconceivable and the only real way to bring online millions or potentially billions

• 7:28Ali Yahya?

  or more AI agents into the kind of the financial system is through a technology that's fully based on software and is as efficient as like the crypto rails that are now available and now can be used.

• 7:39Erik Torenberg?

  Yeah. So say more about some of the use cases that stable coins are currently enabling. Is it mostly on institutional level? Is it on a consumer level or where's the common interactions people are having?

• 7:49Speaker 01

  I think it's both. It depends what markets you're talking about. There's a company in our current accelerated batch called Zarpay, which is operating in Pakistan and they're basically creating a network of small, you know, the little shops. If you've been to Africa or wherever, they have these little sort of mobile kiosks where you can put money on your phone and that sort of stuff. And so they're basically using that network in order to create a way for people to come in and deposit their local currency and get stable coins. And then they're building a whole suite of like financial services around this as the atomic unit. And I think a lot of countries that have unstable currencies or other financial issues for which holding dollars or the equivalent of dollars in stable coins is very appealing, immediately understand the value of this and are very attracted to using it.

• 8:33Speaker 01

  So I think it goes from that all the way through to banks and financial institutions. I think in many cases, there's been an interest in crypto and some of the banks and financial institutions have wanted to get involved, but it's been very unclear how they could do it, largely as a result of the lack of regulatory clarity, but also because crypto can be a little scary or whatever. And so it hasn't always been obvious for them to. See a path, how do we get involved? What's the way that we can bring this to our consumers? And so, stablecoins, I think, are kind of like a baby step in, in the sense that it's a lot more clear what the value proposition is. It's a non-speculative use case. And so, I think it's just a good entry point for some of these larger institutions.

• 9:15Erik Torenberg?

  Yeah. Help us understand better the stablecoin landscape around like, what big companies or types of companies have emerged or will emerge as a result of it, or how it impacts the crypto startup ecosystem more broadly.

• 9:26Ali Yahya?

  So, right now, at the center of all of the action are the stablecoin issuers. So, we've got two of the major ones right now are USDC, which is created by this consortium between Coinbase and Circle. And then, there's Tether. And so, both of these are like the kind of the biggest two issuers of stablecoins today. Then, of course, both of these stablecoins operate on top of blockchains. So, another important piece of the stack is the infrastructure on top of which some of these stablecoins operate on. And then, you have this kind of collection of companies at the periphery that are generally just companies that help connect the crypto world to the external world. And that would include wallets. It would include some of the fintech companies that are using blockchain technology as the backend,

• 10:08Ali Yahya?

  but have a frontend that looks more like a Web2 type of frontend and doesn't expose the crypto aspects to the end user as much. And all those players will be a part of the story as well. So, one of the things that we talk a ton about is, what will this stack look like end to end as the space evolves? And one of the exciting things that we are hoping will happen soon is it will get legislation that sets the rules of the road for stablecoins and for what is required for an issuer to create a stablecoin, what kind of collateral is needed for the stablecoin to be compliant. And what that likely will do is, if that works, and we strongly believe it will likely happen this year, it will, to some extent, commoditize the issuance layer because it'll be easier for new issuers to emerge

• 10:53Ali Yahya?

  and create their own stablecoins that are also USD-denominated, to the point that those new stablecoins are somewhat fungible and interchangeable with USDC and with Tether. Because if all of them are compliant, then you can trust that all of them are likely to be. Ultimately redeemable for a dollar and they're equally trustworthy, which means that then the issuers may no longer be the ones that capture all the value the way that they do now. And instead, a lot of the value might be captured by some of the other layers. Like, for example, the infrastructure is likely to capture a lot of the value because a lot of the activity, a lot of these sort of transactions, stablecoin transactions that are happening, happen on blockchains like Solana, Ethereum, Zui, a number of other kind of important Layer One blockchains.

• 11:35Ali Yahya?

  And all of those require payment of gas for all of those transactions. So those blockchains are likely to be important players in the way that this unfolds. That's on one another stack and then the other end will be kind of the endpoint, the interface that connects this whole crypto world to the end users. So wallets will be likely important. One of our portfolio companies, Phantom, likely be well positioned as a gateway or in an interface for people to interact with stablecoins and get kind of exposure to US dollars regardless of where they may be. So that's maybe a bit of a layout for what the ecosystem looks like at the moment.

• 12:08Erik Torenberg?

  Yeah. And it seems like for yours, there's been this question of, hey, what's going to make it so that there's hundreds of millions of users? I'm not sure what it is at the moment across all of crypto, or a billion users. They asked before that, kind of, what's the iPhone moment? What's the product that everyone's going to be using that's also a platform for everything? Is it stablecoins or is it something else? Or how do we think about that?

• 12:25Speaker 01

  I think the odds are good that stablecoins are that thing. I also don't think that necessarily there needs to be one thing. I think we mentioned AI. Ali has made some investments in that category. We've done some as a team. I think there's going to be different waves that bring in different users. A while ago, Web3 games was a big entry point. Now it's AI and stablecoins. So I think the users do come in waves. I think there's a lot of it that sort of tracks the cycles that we see every couple of years in crypto. Chris always talks about, do you want to be the indie band? Or do you want to play like the Super Bowl or the mega stadium? And I think like stablecoins really have the ability to appeal to like a much broader audience because, like we said, it's just a use case that makes sense. It's pretty clear what the value proposition is.

• 13:09Speaker 01

  And so it appeals to a broader audience.

• 13:11Erik Torenberg?

  Yeah.

• 13:12Ali Yahya?

  Yeah, in part also because it addresses a very real pain point, whether it be for people in Third world countries that want exposure to the dollar because their local currency may not be as reliable, or people who want to move money between borders, and we talked about how that can be extremely inefficient, or even companies that want to move money across borders, they still have to deal with all the inefficiency. Apparently, companies like SpaceX are already using stablecoins for treasury management to move money from one country to another in a way that's much more efficient than using the traditional financial.

• 13:44Speaker 01

  Yeah, I believe they were using Bridge, which Stripe now acquired. Yeah, I mean, it's interesting, Stripe Sessions, their big conference was like all about stablecoins. So many of the talk tracks last week were about that, and so I think it's really indicative of the fact that this is permeating not just crypto companies, but out more broadly. And I think, obviously, the other necessary element in addition to the infrastructure improvements and all of that is just the fact that now we have a more friendly regulatory regime, which is interested in seeing these kinds of things flourish.

• 14:10Erik Torenberg?

  Yeah. Just one meta point, it's one thing I've always appreciated about crypto investing is you guys, as domain experts, don't just need to understand the technology, which is complex enough in and itself, but you also need to understand the policy, regime, law, monetary policy, economics, foreign policy, and how all these things are intersecting with crypto startups.

• 14:30Speaker 01

  Yeah, I mean, I'm certainly not the domain expert on some of the policy stuff, but we've really assembled a super strong team who's been very involved in DC and trying to push the ball forward for the whole industry.

• 14:38Erik Torenberg?

  Yeah. You mentioned Stripe getting deeply involved in crypto. It's interesting, because people often contrast it with AI and say, hey, AI is mostly sustaining innovation, and that, of course, there's massive companies that have been formed, but a lot of the gains have gone to the biggest companies, whereas crypto is mostly a startup, though some bigger companies are getting involved, too. It's funny, like, maybe Facebook was just a few years too early. If they had launched Libra in a more friendly regime, might that have worked? How do you think about the startup versus incumbent distinction in this space?

• 15:06Ali Yahya?

  Yeah, crypto is like a fundamentally radical set of technologies that is very, very hard for incumbent players to adopt and run with, precisely because it is so fundamentally disruptive to the way that they do things. I was at Google a while back, I was at Google X, working on a robotics project, but I was already very interested in crypto. And Google X is supposed to be like the moonshot factory and it's supposed to be super innovative and open to new ideas, open to start new companies.

• 15:33Erik Torenberg?

  But they didn't want it?

• 15:34Ali Yahya?

  New ideas. I tried to get Google X to.

• 15:36Speaker 01

  I tried at Facebook, by the way, same thing.

• 15:38Ali Yahya?

  And it's like, Google would not touch crypto with a thousand-foot pole unless it was like a very, very kind of vanilla, we will run some node or whatever.

• 15:45Erik Torenberg?

  Were they like, this makes no sense? Or were they like, it's evil?

• 15:47Ali Yahya?

  I think, yeah, they kind of fundamentally didn't get it. They were afraid about the optics. They were afraid about the regulatory association with it, the reputational consequences. Also, like, the whole Web 3 vision, the vision of decentralizing web services, which is, I think, the kind of the most futuristic vision for crypto, is fundamentally disruptive to the way that these companies work. Like, these are centralized companies that make money and have power by virtue of being so centralized.

• 16:15Erik Torenberg?

  Yeah.

• 16:15Ali Yahya?

  And if you build something like a social network that is fully decentralized and has no core central company, like no monopolistic tech giant that's worth $44 billion that controls what recommendation algorithm is used and who gets to follow whom, and all of the data and the social graph itself, then that company no longer has a business model.

• 16:37Erik Torenberg?

  Right.

• 16:37Ali Yahya?

  It's a very, very different business model to build a social network that's decentralized in the way that, say, like a company like Farcaster.

• 16:44Erik Torenberg?

  Right.

• 16:44Ali Yahya?

  Currently is. And so for a company like, say, Facebook or Google in its own way to decentralize itself and to truly embrace crypto with arms wide open, it would have to cannibalize its own business model. And I think that's actually becoming true for AI as well. Like, I think that it was very, very true that AI was a very sustaining innovation.

• 17:03Erik Torenberg?

  Right.

• 17:03Ali Yahya?

  Before, but it's gotten so powerful that there are many elements of it that are disruptive. Google wanted to truly embrace AI, it would have to replace search.

• 17:12Speaker 01

  Search, yeah, exactly.

• 17:13Ali Yahya?

  With an LLM instead of having its current model. And of course, that's a hard thing for it to do, given the kind of the insanely profitable business model that they currently have.

• 17:21Speaker 01

  Yeah.

• 17:22Erik Torenberg?

  Yeah, that's fascinating. Is that still the vision, that we'll have decentralized social networks, decentralized marketplace? Or where are we on that vision? And what are the bottlenecks to that vision being realized? Of networks at scale that are truly decentralized and competing with some of the centralized ones? Is it technological or is it that people just don't really care about this in the same way? Or like, why hasn't it happened yet?

• 17:41Speaker 01

  I think it's mostly a consumer preference issue. I do think like now some of the products have gotten really good. Like Farcaster, for example, the product experience is very good. But it's just challenging to get people to switch. Because the reason you're on the social network is for the graph. And so it's difficult to export an entire graph. I think like users are accustomed to being the product. If you're not paying for the product, you are the product. And I think in many cases, consumers are used to that experience. And you know, ads are annoying, but they're not necessarily that bad. And so people just accept it and don't think too much about it. And by the way, this is interesting because if you look at like all of the big social networks, none of them have been started in the last decade.

• 18:25Speaker 01

  And that's not true just of crypto, it's true in general. And so it's just very difficult, I think, to get over this hurdle of reaching a critical mass whereby people actually say, oh, I'm in the network and I'm going to stay in the network. So it's not just a crypto thing. I think it's just difficult nowadays. People only have so much attention and with the networks that there are, most of the attention span has already been captured. So I think we may need to see some of the existing ones falter before there's like enough room for some of the new ones to really take hold. We'll see.

• 18:56Ali Yahya?

  Yeah, we used to believe that these ideas and these companies would be the first to gain adoption. And that was largely because all of the financial use cases, like the DeFi use cases, even like the stablecoin use cases were legal. This was the case in the previous administration. And so it felt to us like the more kind of innocuous seeming social network use cases, gaming use cases would more likely gain adoption. And then that will be the gateway for other things to eventually become legitimate and get acceptance from a regulatory standpoint. But now that the regulatory landscape has shifted so much, to the point at which it's now a much more friendly landscape. And as a result, we have all these traditional financial institutions getting involved and stablecoins are really. Having a moment, combined with the infrastructure clicking into place,

• 19:40Ali Yahya?

  it's now much more clear that the more financial use cases are likely to happen first. Those will act as a legitimizing force for the rest of the space. And then the consumer use cases, which we still believe in, will take longer. And I think, as Aran is saying, it's very, very hard to get those things right. Like, the bar that a consumer has on the quality of a consumer-facing application is extremely high. Yeah. And crypto has not yet figured out all of the UX challenges. And the kind of the seamlessness and usability challenges of crypto are still... It's still a nascent technology on that front. So it'll take longer for all of those things to get resolved. But in the meantime, we have all these other financial use cases, which I think will solidify the technology, will legitimize the space for a broader group of people, will get more entrepreneurs to come into the space. Right.

• 20:23Speaker 01

  Yeah, I think on the point of the attention span, or lack thereof, of consumers, it's interesting when you see a new network created around an area that doesn't already have somebody in the non-Web3 world occupying it. So, for example, I think a good example of this is Blackbird, which is kind of a network for restaurant lovers. And you can think about it as Amex points for restaurants. And they're occupying a space that nobody really owns it right now, like the credit card companies do, but it's still sort of a so-so experience at best. And when you have a great entrepreneur who is really deep in restaurant tech, like Ben Leventhal, who's the founder, tackling a problem like that, bringing a consumer Web2 experience, but using Web3 as the ownership layer,

• 21:08Speaker 01

  therefore allowing the restaurants and the consumers to actually have ownership in this network, which wouldn't be possible in a Web2 context, then it's pretty interesting, because that's something that you couldn't really give the same ownership. And if you look at some of the platforms like Uber Eats or DoorDash, the restaurants have to work with them because their margins are so slim. And so they need as much volume as they can, basically. But it's not great, because the platforms are, in many cases, quite extractive and dig deeper into the margins of the restaurants. And so if you are using stablecoin payments to bring down transaction costs and you're also giving restaurants actual ownership in the network, therefore helping their bottom line in that way, too, it's really interesting.

• 21:48Erik Torenberg?

  Yeah, in order for it to match the consumer expectations that we're talking about, it has to be a new interaction model or new value that's unlocked that goes straight to the consumer as opposed to something that's like abstract, like the same product, but just decentralized. It has to be forecasted with frames and some of its other experiments has been net new things and only it could do and a web two couldn't. And Blackbird is another example of that.

• 22:07Speaker 01

  Yeah, exactly.

• 22:08Erik Torenberg?

  Ali, let's go a bit deeper on AI and the intersection between AI and crypto. What's working there or where are you most excited?

• 22:14Ali Yahya?

  Peter Thiel actually had this tongue in cheek line back in 2018, which I think rings true, which is that AI is communist and crypto is libertarian. I think the meta point is that these two technologies are very different from one another and in many ways they're actually counterweights for each other. So there are many ways in which they are intersecting and we can talk through a few of them.

• 22:38Erik Torenberg?

  Yeah.

• 22:39Ali Yahya?

  I think one of the most important ways is that AI is creating a kind of overabundance of media and of human looking entities, agents that can pretend to be human or deep fakes of video or audio that seems very human and it's hard to know whether you're looking at something that's real or something that's purely generated. And crypto happens to be a really good technology to help authenticate media or help authenticate data in general. One of the ways in which these two worlds will collide is that there are crypto projects that are working on, among many other things, proof of humanity, which would allow someone, anyone, a user on the internet to prove that they actually are human so that anyone

• 23:21Ali Yahya?

  on the other end can know that they're interacting with a human and not with an AI bot or an AI agent. WorldCoin is one of these companies, is one of our portfolio companies, and they built an orb that uses biometric information and also uses zero knowledge proofs to keep all of the biometric data private. In fact, the data itself never leaves the orb and only a code or a cryptographic object that is derived from the biometric data ever leaves the orb. And from that cryptographic object, it is not possible to infer anything about the biometric data itself. It's a technology that allows anyone to prove their humanity on the internet. There was that famous line in the 90s that on the internet, nobody knows you're a dog.

• 23:58Erik Torenberg?

  Yeah.

• 23:59Ali Yahya?

  And that is very, very true now in 2025 where like on the internet, nobody knows you're human. Like you could be anything, you could be.

• 24:05Speaker 01

  A dog or an ape.

• 24:06Ali Yahya?

  A dog or anything else.

• 24:07Speaker 01

  Yeah.

• 24:08Ali Yahya?

  So that's one way. I think that cryptography, blockchains will help deal with the immensity and abundance of signal and noise that AI will generate. Another big one is that crypto can help decentralize the power structures that are emerging in AI. At the moment, it seems like there will be a small number of very, very powerful players in the AI world. Even though it's unclear as to whether there are things like network effects that drive the sensibility, there are just a handful of really powerful players in the space, at least at the kind of the model layer, like OpenAI and the other kind of big companies that build foundation models. Crypto offers an alternative for creating AI systems that's more decentralized.

• 24:50Ali Yahya?

  And so an example of this is a company called Jensen, which is also in our portfolio, that builds a kind of marketplace for compute. So someone on one side of the marketplace can provide their idle GPU capacity to the network, and then someone on the other side, who might want to use the GPU compute for training a model or for doing inference on a model, can, through the network, make use of all of that compute. And the network manages all of these heterogeneous computational resources to create something that feels like a unified cloud, on which you can run all of these machine learning AI workloads, in a way that's fully decentralized, in a way that is not controlled by a single company, and in a way that could actually be more efficient than a cloud by virtue

• 25:33Ali Yahya?

  of using capacity that otherwise would just go idle and unused. It's just capacity that's locked away in all of these like pockets that are removed and not in one particular data center. There are many hard technical challenges to get there, but there are a lot of smart people working toward figuring that out, and we're very optimistic that'll also happen. It'll also allow for machine learning workloads to run in a way that is also verifiable. So this way, you don't have to trust a centralized company, say Facebook, or like one of the social media companies, that the machine learning model, that the AI model that they're running, for say, like their recommendation algorithm is unbiased or has particular properties, you can actually, with cryptography, verify that those things are the case and

• 26:18Ali Yahya?

  that these things are executed in a way that's correct. You can use some of these decentralized systems for that as well. And then maybe the final one, which I think is the most futuristic and the most challenging, is having crypto help AI figure out the new business models for the internet.

• 26:34Ali Yahya?

  So one of the issues that AI will create with the current business models of the internet is that right now, the way that the internet works is that you have an aggregator, like a search engine, driving traffic to creators of media, like say, someone who has written a blog post, or someone who has like a page that has content. And there's ads as like the business model that mediates that whole interaction. And that entire business model goes away if you just have an AI that just gives you the answer that you're looking for. So instead of doing search on Google, getting exposed to a bunch of ads, and then clicking through to a website, and having all of those parties be happy because a business model includes all of them.

• 27:17Ali Yahya?

  Now, instead of that, you just interact with an LLM, you get the answer immediately, and you never click through to the final page, and you never get exposed to an ad. That kind of completely changes the way that the internet works, and we're going to need new business models for the internet if that's the case. So one idea is that you could through, there are a lot of these research efforts to try to figure out attribution in the training of a machine learning model for what pieces of data contributed to a particular output. So you're asking an LLM a question, you kind of want to know what pieces of data that were used to train that LLM contributed to the answer that the LLM ultimately gives you when you ask it a question.

• 28:00Ali Yahya?

  And if you could know that, then you could come up with a business model that rewards the people who contributed that data originally. And crypto could be part of that. So there are open problems on both sides, like you have to figure out this attribution challenge in the AI world, and there are people working on that problem. And then there's a challenge on the crypto side, like how do you build a network that can, using that information, compensate all of the parties involved in having the AI actually give you what you ultimately want?

• 28:26Erik Torenberg?

  Fascinating. Are the big labs interested in crypto? Do they need to be interested in crypto for this to happen? Or is this largely coming from startups? It's funny, Sam Altman, of course, OpenAI, but also WorldCoin, so it has some familiarity. But what can you say about this?

• 28:38Ali Yahya?

  For the most part, I don't think so. I think for the most part, the AI labs, they're just running with AI, and there's so much that's exciting in that world that I think crypto doesn't really factor in at all. But there are crypto companies that are very interested in AI and are thinking about the ways in which crypto will ultimately... Make a difference in that world.

• 28:56Erik Torenberg?

  Yeah.

• 28:56Ali Yahya?

  So like this company I mentioned Jensen, actually the founders are very, very deep in AI. They have an AI background, but they also happen to have a deep commitment to building things as open source networks that are ultimately decentralized. And so they are of the few people who really do straddle both worlds.

• 29:13Erik Torenberg?

  Zooming out a little bit, what are some of the biggest misconceptions you think people have about the space right now?

• 29:18Speaker 01

  Well, I mean, I think for the last few years, it's been really challenging to launch a token network in the United States in particular, because there was a lack of clear legislation. And then there were very aggressive folks in several agencies working on basically not allowing entrepreneurs to launch networks. And that applied to entrepreneurs who were very well meaning and very much wanted to do things by the books. And so I think one of the challenges was that people obviously didn't want to end up in legal trouble and therefore, in many cases, pulled back their plans on that front, which really impeded their progress on the product side as well. So they couldn't really build their vision, because I think tokens are part and parcel

• 30:03Speaker 01

  of what's valuable and interesting about crypto. And so if you remove that piece, it doesn't make any sense. So the thing that's a misconception perhaps is that the situation is very different now. We have a much friendlier administration in place. We have a very different situation in terms of the leadership of these agencies now. And so I actually think it's a great time for folks to be building token networks. And I think that message hasn't necessarily fully made it out there. So I'm hopeful that more entrepreneurs realize that the situation is, again, very different from what it was just a few months ago and start to come back in force.

• 30:38Ali Yahya?

  I completely agree with that. I think another big one is outside of like our immediate circles, like outside of the world of tech, it's shocking to me that people continue to think of crypto as just like a thing that's supposed to be money only, or they think of blockchain as a kind of ledger for money. And I think that that misconception comes from Bitcoin, from Bitcoin trying to be money and only money and not really trying to be anything else. And the fact that This misconception is that Ethereum is like Bitcoin, and Ethereum is actually the silver to Bitcoin's gold. And that all that crypto really is is just another kind of attempt to doing what Bitcoin did.

• 31:18Ali Yahya?

  The fact that Ethereum actually is a fundamentally different thing than Bitcoin is still not widely understood. The fact that Ethereum is actually a kind of computer, where you can build all sorts of different applications, where the software that runs on top of the computer has unique properties that no other software has ever had, is, I think, not widely understood. And these are programs that, like the programs that run on a blockchain like Ethereum, are programs that have a life of their own. They are programs that can make commitments that no one has to trust anyone to believe in. It's a program that is essentially free from interference from anyone, including the people who originally wrote the program.

• 32:00Ali Yahya?

  And so that's a very unique property that no other kind of software has. It's a kind of technology that inverts the power relationship between the software and the hardware. Whereas historically, the hardware has always had power over the software, because whomever controls the hardware can turn off the software or change it in some way. Whereas in crypto with blockchains, the hardware commodities, these are people who run, like the miners, for example, or validators in the blockchain context, don't have any power over the software that runs on top. And that's what makes a blockchain unique. And that's what makes it capable of doing so much more than just money. You can kind of build far more sophisticated primitives. So stable coins are the first thing, but the things that come after, things like DeFi, where you can build much more sophisticated financial primitives on chain,

• 32:45Ali Yahya?

  or some of these other more futuristic ideas, where you can do AI, you can do deep end, you can do some of these consumer-facing applications, like decentralized social networks. All of that relies on the properties of a blockchain computer. That's not just a ledger, it's a full-on computer on which you can build applications. That, I think, is not something that most people really get.

• 33:06Erik Torenberg?

  Yeah. Maybe gearing towards closing here, Ali, can you give a bit of an update on kind of the smart contract platform wars as an outsider or someone who's paid attention at certain times and not at certain times? What I've heard or gleaned is Bitcoin, as you mentioned, has tried to be money, but there's a little bit of a nascent Bitcoin builder movement. I'm not sure if that's led to something particularly meaningful in the space. And then my understanding is that Ethereum has tried to optimize across multiple dimensions, both trying to be money, but also trying to be the base layer for sort of decentralized internet and committed to decentralization in a way that some people think is at the sacrifice of usability, whereas Solana has not had the same commitments to decentralization, has really optimized for usability. One, is that a fair characterization? How would you edit the characterization? And two, how has this all played out or where are we right now on that level?

• 33:49Ali Yahya?

  That's actually a really good characterization. The way that I would break things down is that there is a very large and multi-dimensional trade-off space, and it's very hard for any one system to cover the entire space. So it makes sense that you'd end up with different systems specialized for different things and then as a result, having different use cases and different value propositions. So Bitcoin, I think, has been extremely successful at becoming like a kind of digital gold. It's been extremely volatile, but I think that there is this belief, there's a this memetic value that Bitcoin is long-term, a pretty good store of value that will be around for a very, very long time. It's not going anywhere and will have properties that are desirable that are not provided by other things like fiat or gold itself or anything.

• 34:34Erik Torenberg?

  It's funny, it's only been around for less than 20 years, but in my head, I treat it as gold. It's going to be there forever.

• 34:38Ali Yahya?

  Exactly. Exactly. So it's really succeeded at that. And I think some of the things that has helped it succeed at that is the fact that it is so hard to change and the fact that it is so simple and you can't do that much with it. Those things are disadvantages in some contexts, but they're real advantages when trying to solve for that particular thing. Then there are like all of the other smart contract platforms that are trying to do much more and are trying to be computers. And Ethereum lands in some part of the trade-off space here where they do really optimize for decentralization and they are fully decentralized. And so it's hard for Ethereum to change quickly because there are a lot of stakeholders and a lot of people who want to be able to influence its direction. And so the choices that it has made have made it a pretty good platform for some of

• 35:20Ali Yahya?

  the higher stakes, like DeFi applications, or for the issuance, for example, issuance of new assets on Ethereum. That might be the default, simply because it's been around the longest and its high amount of decentralization make it very suitable for that. And then there are blockchains like, say, Solana and Sui, which are extremely high performance. They're very well suited for transactions and for payments and for things that do require that level of performance. If you wanted to build something like the NASDAQ exchange on chain, there's no way you're doing that on Ethereum 0.1. You probably need a blockchain that has the kind of the level of performance that a Solana or a Sui or some of the other kind of more modern or more recent blockchains have. So I think I expect that each of these ecosystems will likely find their niche.

• 36:01Ali Yahya?

  It's obviously very uncertain and there's all this talk about how maybe Solana will eat Ethereum's lunch and that's a possibility.

• 36:09Erik Torenberg?

  But it's still wide open, is basically what you're saying?

• 36:11Ali Yahya?

  Yeah, it's wide open and there's like a lot of ways in which it could play out.

• 36:14Erik Torenberg?

  Yeah. Closing out, Arianna, I want to double click on your point about the misconception in terms of how the policy regime has changed.

• 36:20Speaker 01

  I mean, I think if you look at the Novi, Libra, you know, it had seven different names, so whichever one you want to use, that was something that could have been incredibly interesting because you have Facebook now Meta with such an enormous distribution network, already has all the users. Integrating payments into that via crypto made all the sense in the world, but obviously they were told in no uncertain terms that that was not something that they could proceed with. And then unfortunately, the whole project died. I will say it went on to flourish in other forms because we're investors in Missin and Sui, yeah, they spun out of there. So there have been actually a number of great teams who came from there. So I think the diaspora of talent has continued to fight the good fight and build.

• 37:03Speaker 01

  But in general, that's another project that I think as it was initially conceived, had to die on the vine because of that. So I think as investors, it's not necessarily our job to envision like what is possible, but rather to recognize it when we see it. And so I'm personally very excited to see what entrepreneurs come up with in the next couple of years now that we have a new opportunity space.

• 37:24Erik Torenberg?

  That's a perfect place to wrap. Ali, Arianna, thanks so much for coming to the podcast.

• 37:27Speaker 01

  Thanks, Eric.

• 37:28Ali Yahya?

  Appreciate it.

• 37:31Speaker 02

  Thanks for listening to the A16Z Podcast. If you enjoyed the episode, let us know by leaving a review at ratethispodcast.com/A16Z. We've got more great conversations coming your way. See you next time.