Epicenter - Learn about Crypto, Blockchain, Ethereum, Bitcoin and Distributed Technologies - Fabric Ventures' Investment Thesis Since the Dawn of Bitcoin to AI - Richard Muirhead

Episode Date: June 28, 2024

It is one thing to invest in early stage companies, but to do so in a nascent industry that constantly reinvents and rediscovers itself is a whole different venture. Richard Muirhead co-founded Fabric Ventures in 2012 around the thesis of supporting the Open Economy. While at first this meant investing in Bitcoin-related projects, once Ethereum was announced, a new horizon unveiled, filled with tremendous potential. Since then, the Open Economy thesis was adapted to match technological breakthroughs, and it is now centered around artificial intelligence, distributed computing and self sovereignty (over both fungible, as well non-fungible assets).Topics covered in this episode:Richard’s background and how he founded a crypto VCFabric Ventures’ thesisHow the AI thesis evolvedUser-generated AIOn-chain autonomous agentsUpcoming DeFi innovationsWeb3 trends that took Richard by surpriseMissed opportunitiesThe impact of crypto liquidity & market cycles on VC investingHow to vet projects in early investment phaseEpisode links:Richard Muirhead on TwitterFabric Ventures on TwitterFabric VenturesSponsors:Gnosis: Gnosis builds decentralized infrastructure for the Ethereum ecosystem, since 2015. This year marks the launch of Gnosis Pay— the world's first Decentralized Payment Network. Get started today at - gnosis.ioChorus One: Chorus One is one of the largest node operators worldwide, supporting more than 100,000 delegators, across 45 networks. The recently launched OPUS allows staking up to 8,000 ETH in a single transaction. Enjoy the highest yields and institutional grade security at - chorus.oneThis episode is hosted by Sebastien Couture & Friederike Ernst.

Bitcoin is digital gold. That was actually the way we're thinking about it at the time. Deploy in Bitcoin and you'll do well kind of thing. We were fully open to the opportunity of decentralized finance. In a sense, Defy hopefully will become useful by being applied as a primitive, as a building block, to the applications and marketplaces we're trying to build. And then in that way, we'll see it become much bigger. The open movement, the ability to share data, the ability to share research, That has been the origin.

That has been the petri dish from which all the notable inventions have come. I think that the other side of the impact of AI on Web 3, just a very specific one, is that if we look at the way in which AI co-pilots are tracking and their ability to create applications, that we're going to go from a situation where there are relatively few solidity programmers and it's incredibly hard to audit smart contracts to a world where this episode is brought to you by NOSIS.

NOSIS builds decentralized infrastructure for the Ethereum ecosystem. With a rich history dating back to 2015 and products like Safe, CowSwap, or NOSIS chain, NOSIS combines needs-driven development with deep technical expertise. This year marks the launch of NOSIS pay, the world's first decentralized payment network. With the Gnosis card, you can spend self-custody crypto at any visa-accepting merchant around the world.

If you're an individual looking to live more on-chain or a business looking to white-label the stack, visit nosispay.com. There are lots of ways you can join the Nosis journey. Drop in the NOSIS Dow governance form, become a NOSIS validator with a single GNO token and low-cost hardware, or deploy your product on the EVM-compatible

and highly decentralized nosus chain. Get started today at nosis.io. Kars1 is one of the biggest node operators globally and help you stake your tokens on 45 plus networks like Ethereum, Cosmos, Celestia, and DYDX. More than 100,000 delegates stake with KORS1, including institutions like BitGo and Ledger. Staking with Kors1 not only gets you the

highest years, but also the most robust security practices and infrastructure that are usually exclusive for institutions. You can stake directly to Quarice 1's public note from your wallet, set up a white table note, or use the recently launched product, Opus, to stake up to 8,000 eth in a single transaction. You can even offer high-year staking to your own customers using their API. Your assets always remain in your custody, so you can have complete PC. of mind. Star Saking today at Coros.1. Welcome to Epicenter, the show which talks about the technologies, projects, and people

driving decentralization and the global blockchain revolution. I'm Sebast Sankuizio, and I'm here with my co-host, Fredike Ernst. Today, we're talking with Richard Murahead, who is co-founder of Fabric Ventures and also on the Near Foundation Council. So Fabric has been a long time VC in the Web 3 in crypto industry, and it's been a long time coming since we should have had Richard on the show. But, you know, 10 and a half years later, here he is to share Fabric's latest thesis, which is about the intersection of crypto A&I, and we'll also get in some other topics like investing in crypto. And, you know, as an emerging manager,

I'm also very interested in asking Richard a bunch of questions about how he runs this fund. So this is also sort of a good opportunity here to learn about crypto investing. Richard, thanks for joining us on the show this week. Fantastic to be here. Don't tell anybody that took me 10 and a half years to get on the show because they might make a jab at my punctuality. And also, yeah, she's something you reminded me, said just then, sorry, reminded me of Harry Stebbings with his 20-minute BC show

where he has done an incredible job of getting great guests on and talking to them about the art of. of venture investing and built that into a great franchise. So I wouldn't be shy of that. That seems like a fantastic way to go, not that I necessarily have anything to add to the picture. And also, I almost appeared on his show right at the very beginning, but it didn't happen. And so maybe I'll have to wait until he's on his 10th year as well before I get to do that. Well, let's hope not. I mean, actually, I mean, that's sort of not entirely true because you were on our 10-year anniversary show. You appeared there for a brief moment, but it wasn't like a proper

interview. But as I understood, as I've learned, I guess, like fairly recently, actually, you've been listening to Episinar since pretty much since the beginning, right? Yeah, somewhere, I mean, I was, I looked back at some of the episodes to kind of, sort of reacquaint myself, but I know that I kind of got the bug somewhere, we got sucked down the rabbit hole or the wormhole as we actually like to call it here at fabric in the kind of spring of 2013 having kind of spent a little a little while building up to that but I know that I then got pulled into kind of running one of my portfolio companies in 2014 and I don't know somewhere between the kind of the autumn and the spring epicenter became a

sort of staple part of my kind of stress relieving sort of jogging routine and reminding myself that I could get back to the kind of romantic glories of decentralization and Bitcoin at some point, which I finally did. Yeah, cool. And I think a lot of people were listening to Epicenter while running back then, including myself, I'd re-listen to episodes sometimes like while going out for runs, when I used to run, which is something I haven't done in a long time. But yeah, maybe let's, you know, for folks who don't know you, dive into your background a little bit,

So, I mean, Fabric has been around, I think, since 2015, 2016. What were you doing before that? And how did you get involved in, I mean, how did you start a crypto fund, basically? The moment of kind of self-indulgence to tell the classic crypto kind of origin story, but I'll try to keep it on point and entertaining. And, yeah, I've been in it long enough that I've heard a lot of crypto origin stories. So I did engineering at university and kind of always wanted us. start a company of some description.

And I kind of fell in love with software, the general power of software. I mean, I tried my hand at it when I was kind of, you know, 9, 10, 11, 12 years, I think as many folks do. Discovered, actually got my first computer from Clive Sinclair, if you know who he is, his parents were neighbors. They got the ZX81 back in kind of 1981. Anyway, he's famous in England, at least, as one of the early. I, you know, major protagonist stakeholders in the PC revolution.

But I quickly discovered that I was not of the, I didn't have strong aptitude for software development itself. And I came out of university and did some strategy consulting, part of which was in the telecom space. And essentially, along with my brother Charlie had an epiphany that there was immense power in the openness and permissible. missionless nature of the IP protocol, internet protocol, data networks that were rising up in the kind of early 90s, but that if we were going to use them to run the world's economy, that they,

you know, maybe need to be fixed to things like security and quality of service and so forth. And actually with the advice as someone who turned out to be a long-term collaborator, Stephen Waterhouse, seven, we, who was an advisor. We built a software company to run these very large scale IP data networks. And critically, I think in this question of kind of getting into the crypto decentralization space, we were quite early, it turns out, with using open source software kind of in anger at an enterprise or at a kind of a carrier scale. So we used an open source instantiation of Corber actually. It's called Omni Orb.

it was a published subscribe kind of middleware and we use that in order to build this product and ultimately IPOed London Stock Exchange and then NASDAQ and they did a merger and then sold to Oracle and that software is still being used today actually to manage large IP networks

and came out of that had a brief that sort of got a taste and this is related again to this question of getting into venture I got seduced somewhat to spending some time with Axel Parna's who are obviously pretty storied and famous and infamous these days, not these for Facebook.

Back now 23 years ago when they were setting up their London office, actually it came about, and I found interesting, I actually re-energized this connection just last week at an event. I basically ended up ordering a margarita at a bar with Rob Blazer who ran Real Networks, of those of you who remember real networks from last century and it's still running today. And Jim Breyer from Axel Partners, as it turns out, and he just said, hey, you should meet my partner, Kevin Camerle, who's setting up in London.

And for me, interestingly, that's one of those fortuitous connections we should always be kind of striving to kind of make. I incubated a company there at Axel Partners, both getting an understanding of what it takes to kind of build a new company from scratch, the kind of somewhat daunting task of looking into the blank sheet of paper, but also learning about the Silicon Valley flavor of venture investing, which I think is still something that we shouldn't necessarily just be looking to ape or copy in the rest of the world. We should be plowing our own sort of furrow, but of course we can be inspired by. But also built a second company that used machine learning primitives to augment

the function of quite sort of knowledge workers operating actually in the IT management space. And one of the things we encountered there was a kind of coordination problem, incentivization and coordination problem. So how to get people to give up their data, you know, they're kind of hard-won data for how to operate a particular part of the kind of back office of a bank, for example, when actually that very same data was the way they were kind of clinging onto their job because it protected their kind of little fiefdom they had. And so through those, I know those threads are kind of fully apparent,

you know, early work with open source software, looking at machine learning privatives before they became famous with Alex NET and Imogenet and so forth at the beginning of last decade. And looking at incentive structures and collaboration of people, when I stumbled into Bitcoin and I genuinely don't know who was, but I met somebody who was involved in Bitcoin as early as January 2009, but I really don't know who that was. I can remember the type of person who was in Switzerland. And then it was re-energized as a conversation by my friend and colleague, Stephen Waterhouse, aka 7, in the spring of 2013. And for me, just very quickly, the kind of the shoe dropped, not just, you know, should we say into their native money,

you know, not just the kind of extensibility, arguably the possibilities of the blockchain. But also, I can remember particularly really getting a. excited by the concept of instantiating, you know, if not laws, but organizational principles into code in what were called Dax and then became DOWs in 2013. Then in terms of investing, I kind of somewhat reluctantly hung out my kind of my proverbial soccer boots sometime 2009, 2010 when I sold my second company to BMC, and then decided Agile investing and was looking for a thesis that was felt sufficiently impactful, in some sense, is sufficiently crazy, that it really kind of, you know, just like work. And so that's what Bitcoin, blockchain, you know, crypto writ large, web,

Web3, open web, let's, we may we'll get into it. That's what they became, and that's kind of the genesis of fabric. Were you inclined to kind of start another company as a founder rather than kind of a fund? Oh, that gave me severe heartbreak. the concept of not starting another company again. And in some senses, I have successfully scratched that itch by building fabric over these last years. But in some other senses, it probably gets in the way of me focusing just on the, in the vesting.

And certainly we have experimented a lot with quite hands-on activities, you know, researching what's going on, hacking around here and there. some of the kind of products, mining, staking, validating, nothing that we have scaled, particularly within fabric, but we try and keep our hand in in that regard. So I think, but yeah, but it was a bit of a heartache definitely to kind of not, not build another company. Because whatever you do, and look at, you know, I did not have household name success. I had some okay success, but I felt like I had way more kind of like tired tracks in my back

and scars on my, you know, scar tissue from things that had not quite gone the way I wanted them to, then I had to as a successor. So you're like, oh, next time I'm going to really do it just right and we're going to kind of get escape velocity. But yeah, it's a good question, Frederick Kim. Richard, can we talk about your fun thesis for a bit? So, I mean, you started a number of years back and it kind of changed over time. So maybe tell us where you started and where you're at now. Yeah, absolutely. So, I mean, inevitably, these things get a little blurred over time and you tend to have kind of the clarity of hindsight. But I'll try to avoid that as much as possible. I kind of mentioned that we were looking for a thesis. I found to start a venture firm in Europe was quite challenging if you didn't have tens of millions of your own money to deploy in it back sort of 10, 15 years ago. So having, I used my kind of company starting kind of instincts that having a focus, a distinctive specialization. and trying to sort of catch a wave would be a good way to tackle it.

So it was kind of on the hunt for something. And I had, as I also mentioned in the back of my mind, some of the principles of digital money, should we call it, some of the power of getting people to collaborate and share their data. But I had not looked too deeply at what was going on with Bitcoin. And then I was exchanging messages with seven in the spring of 2013, you know, asking, and he was over in Silicon Valley and I was in South Kensington, basically, in London. And I was like, you know, what are you up to over there?

And there were two things he said, he said he was looking at a lot of VR and Bitcoin. And so as a result of that conversation, I sort of dove in. And blockchain.com was obviously a big company, you know, over here that was kind of active. And that was one of my first ports of call. And it was made safe. And there was a whole bunch of other people who were active. And we ended up working together on the investment in BitStamp in the autumn of that. year. But in terms of the thesis, when we were thinking about how to benefit, you know, to deploy

capital to benefit from the space, there was still a little bit of a backstop instinct that, you know, look, Bitcoin is digital gold. That was actually the way we're thinking about it at the time, deploy in Bitcoin, and you'll do well, kind of thing. And actually, of course, it'd be interesting, I mean, we're talking like between a price of some $98 to $400, you would have done pretty, pretty well in comparison to energy and funds. Of course that is partly hindsight because you can't say, oh, you should have just done that because that would be a very concentrated

bet to take with no diversification. And on that point, it did feel at that time that there was I tried to conceive of all of the different things you might invest in, the projects to invest in, in order to build a portfolio as an

investor, and it felt relatively limited at that point in time. And I think, of course, it really took January February 2014 when kind of Gav we were co-hosted an event with CoinSgram

and he announced he presented Ethereum for the first time and then you started seeing this obviously the shift to kind of generalize smart contract platforms and then being able to be built on top of that and obviously the Defi world we know that started to make it look like the thesis

you could have a thesis just around that space I think then of course, and this is one of the things I think everybody in this space has to wrestle with. When the price is up, everybody is massively positive and often over-exuberant and thinks that anything is possible. And of course, actually, I find ironically, it's at that point in time that limited partners

or folks who are a bit detached from the space tend to get jolted into action to look at investing. Of course, it is ironic because of course, maybe it's when the price is not so high that you want to be preparing to deploy. And, you know, we went through that winter of 1415 when everybody sort of started concocting the kind of, it's not Bitcoin, it's blockchain kind of narrative

and going a little bit sort of B2B, let's look at how that can be deployed in companies like R3, looking at the kind of the back office for banks and so forth. I think whilst we definitely did, didn't dismiss that. We remained believers in the openness, permissionlessness and so forth of the public blockchain movement. It actually reminded me of when, I would have been like 96, 97 or something. There was a chief scientist of UNET, one of the powerful Internet service

providers at the time. I remember spending some time with him. And he was collaborating with IBM and IBM was trying to get everybody to build sort of private intranets to solve, you know, supply chain problems and to share information between, you know, with GE and Ford and all of the kind of clients. And it was the equivalent of a kind of permissioned blockchain. And those projects didn't go anywhere. Yeah. I mean, ultimately, there was the adoption of the underlying technologies, but those projects went nowhere. And it's, and it's, and so I think for me, it, it, it echoed that. And so we can remained kind of on the kind of public direction. So both Max, Mercia, then Julian Tevenov, those forthcoming years joined. And Julia in particular was very early looking at the

defy space and looking at the arbitrage of opportunities before between exchanges, for example, prior to that, something that occurred to us, but, you know, other people exploited or took advantage or profited from in a way that I certainly never did. And so we were, fully open to the opportunity of decentralized finance. I think it's an interesting question where we've reached with the maturity of that. I think we may be at some kind of local top, but I think it's really a component of something much bigger. I guess the only other thing worth mentioning in terms of our thesis

is that just because of the nature of where it come from in terms of building companies that try to harness data to, gathered from people in kind of networks. We always saw ultimately Bitcoin blockchain decentralized data structures, the coordinating power thereof of tokens that are native to them as a way of organizing the world's data from the bottom up. The emergent power of properties of those networks

are the powerful thing, that if we're going to have a chance of getting the right data to the right algorithm at the right time in the right format so that we can really all benefit to the maximum from what we now see accelerating in the world of AI, then there's a beautiful marriage between those two movements. And that was for a long time in a court of what we pitched and it's gone in and out of favor, but it suddenly seems to be much more in favor today. So let's talk about the AI thesis. As we were preparing for this, I was reminded that, in like 2017-18, there was a moment where blockchain and AI, that narrative had emerged amongst

all the other narratives at the time, Defi, this enterprise blockchain idea that was floating around in the 2015s to, I guess, like 1718 era. And I remember at the time, I remember that narrative just feeling very cringe to a lot of us folks building in defy and permissionless networks and we had people pitch us ideas for the podcast and maybe sponsors and stuff like people reaching out to us to be on the show. You know, sort of pitching this idea of like AI and blockchain or the perfect technologies. But it just felt so cringe at the time. And to me, at least, I didn't see any sort of promise or utility in those two technologies merging and somehow becoming complimentary for each other. What's different now and why is it why is this narrative

now making a resurgence and what what sort of specific technological advancements perhaps have now made it actually interesting and something worth looking at? Yeah. So so I guess one of the things I draw from that actually is that that we all battle you know markets high, market's low with the emotions attached to that. We also battle with the fact that you know trying to sort of actually just stick to what we believe might be right and trying to ignore the other people the opinions of other people, which might be just fighting that instinct of

you know, shying away from what feels a little bit of, you know, cringy, should we say? Although that being said, of course, you know, and I mean, we probably wouldn't get into it, but AI's got a long history when it, the kind of the dawn of the AIIH has been heralded and it's turned out to be a kind of

a false horizon, you know, going back decades, of course, And I think that, you know, it's pretty well documented. One of the things that is different, has turned out to be different now, is that, you know, just the computational power and the quantity of data available to train then some innovations in, you know, the types of AI being generally used, obviously, notably the transformer style, you know, LLMs. They, um, that, um, that, that, That has really kind of be taking things to the next level in terms of the wow factor for anybody using GP3 level at GPT4. So obviously that is that mainstreaming makes it different on the AI side. I think in terms of the kind of the connection between the two, I think first starting with the kind of Web3 side of things, you know, I think there's often in a sense a similar challenge later that the feet of Web3, which is like when are we going to have mainstream apps? When's that going to happen? And I often sort of get into and go, hold on a second, can we just calibrate, you know, how long it's actually been and how far we've come?

You know, just take a measure how many DAP's being built, how many developers being involved, how sophisticated a Web3 stack that's being built and say that's part of $3 trillion in market cap that has kind of been sort of generated in value. that feels like a pretty impressive. And, you know, that I think that work that has gone in to build a Web3 stack that truly can scale to support, should we say, native on-chain payments, to support sort of consumer-grade applications, that has the levels of abstraction in it so that you don't have to care about which L2 you're on, or indeed, preferably which L1 you're on, you're on and or anything about specific to the naming service of, you know, the aforementioned L1, you know, I think we forget about how much has happened. And I think we do, it does feel like we're at a tipping point there. And then I think that all of the work that's gone into creating the defy ecosystem, we always thought about those as kind of necessary primitives or building blocks in that overall stack. You know, this is, you know, this is,

one, you know, this is the first time we've had a wave of software that has the ability to make and keep financial promises, which sounds like a pretty powerful, you know, tool. And, and I think that's where the marriage with AI comes, comes about, because we've seen, and I think really comes about because AI, we've seen how the choke point for that has been the capital required, not so much the electricity consumed, but the sheer capital required to buy the GPUs and get them in a timely fashion in order to train models. There are all sorts of chalk points around electricity distribution and even generation, if you want to have that, is sort of native geotography that are emerging. But I think it's going to turn down to data. And that, you know, a lot of

lot of the kind of open-sourced and commonly available data sets that are of the scale required for LLMs has already been consumed, but there's a huge amount that is still trapped in, for example, social media. And I think famously is just, you know, Zuckerberg said, you know, we're not training Lama 3 on our social data. But if we can find a way in which we can respect people's privacy and perhaps even reward people, instead of people and reward them, for sharing that data. There's a huge treasure trove of opportunity. And I think we can then obviously go beyond social network data to health data and the like and maybe deeper personal financial data. And so I think that's the opportunity now, the maturity of the stack now that is there.

Then it's the opportunity to use those, you know, the capabilities of that stack to go and use token economics to incentivize the sharing of that data. I think one of the most interesting applications for crypto and AI, at least the one that I find the most useful, is this idea of user-owned AI, where blockchains effectively act as a way for users to own their data, to leverage their data if they wish to sell their data to to organizations and folks building models. I think this notion is kind of also emerged previously in crypto when talking about social networks and creating data networks effectively. I think that's one of the most interesting sort of intersection of crypto and AI.

Aside from the obvious marketplace for compute, right, and marketplace for GPUs, which is an sort of obvious use case. In the case of user-owned AI, given how we've seen Web 2 aggregate to very large players, and those have become very powerful, and specifically in the AI vertical, have become very powerful,

do you think that it's even possible or conceivable that we have user-generated AI, or user owned AI in 10 years, that that is even a significant enough portion of the AI market to even matter? I hope it's possible. It's been characterized by Michael Casey and his co-author as kind of our greatest fight.

I think it is a recent book. And look, if we go back and look at it, it's clear that if you follow the money, the natural concentrating tendency of capitalism, we've seen very evident in Web 2, and that this is already, you know, is clearly further exaggerated with the kind of prevalence of or the rise of AI.

And like, it's just driven by, we've seen it time and again, that there are, okay, step back a bit. The open movement, the ability to share data, the ability to share research, that has been the origin, that has been the petri dish from which all the notable inventions have come. But what happens, unfortunately, obviously, and we can observe this, is that there are a few

milestones reached in that research and the aggregation of using the benefit of that data, and then people close off their models. I think just recently, Francois Chalet from Google Web as far as to say that, you know, along the lines of Open AI is put back frontier research in AI by, several years, three, four, five years, since it closed down through its actions, the publishing of that research, and also by virtue of the fact that it has popularized out of them so much that it's sucking the oxygen away from other forms of research. So I think we need to remember the power of openness that's there, that if we're not careful, this concentrating force,

you know, you know, steals from us. And I think that at every level, at every level, you know, level of the stack, and we would think of the stack as, you know, so you say compute data, models, talent and governance. You know, at every layer, there is the potential to use a token economic driven model, perhaps combined with a new form of open source license for the models and the data that has been generated in order to give everybody a share. However, that's easy to coin a phrase somewhat as user end AI. but there are many components to make that that happen obviously.

And, you know, I think at each layer you mentioned, you know, clearly we'll see, you know, Airbnb for GPUs. And we're seeing, obviously, there were great companies like, you know, Jensen and protagonists like Acash and so forth that are out there kind of building things. There are real challenges there, of course, of getting, and it's always, I was thinking of this a little bit as the martini problem. I think it's, I think it was a 70s martini out as like any time, any place. anywhere or something, presentation of a martini.

But anyway, it's the, how do you get the right type of compute power available at the right price in the right spot in the network, at the right time to do what it needs to, you know, to do? That's some of the challenges because there's the difference, obviously, between training, for example, high compute and inference, you know, a requirement for low latency. You know, that matching algorithm of supply and demand is very challenging. in the same applies on, you know, on the data set. How do you, and how do you get that to work? And so that's what people are kind of wrestling with. But I think this is a watershed moment.

Either we get a bit of a vicious cycle where everything gets extremely concentrated, which I think there are very strong arguments philosophically is a bad thing. Power corrupts, absolutely, corrupts absolutely and without getting into kind of you know politics and the last sort of of century or so and then secondly so we can believe in philosophically that's the right thing do but secondly I think we can actually believe that and I've been trying to think of the best way to kind of articulate this but because I think we've we've not seen it yet sort of went out we were an early investor for example in Ocean Protocol which was looking to kind of marry, you know, create a data marketplace. And one of the key things is that there is

kind of no easy marketplace for a large marketplace for all types of data. It's always about kind of very specific data in a specific tiny, you know, more applied marketplace where the value is. And then you have to be able to negotiate the value of the data at that point in time. And a lot of companies that build software effectively are doing that negotiation on behalf of that data set. And so we need to create something that's similar to that allows for those collectives that own that data to negotiate on the fly or the promise of the value that data in advance. I think that's an area of research and of invention, I think, kind of for founders. And what that is in my mind's eye is a rather than, this is a crude metaphor, but rather than saying, we're going to concentrate everything around the Ford Motor Company,

and there's going to be one type of car, and by the way, you can have any color as long as it's black, and the mass production engine is going to be fed, and the profits are going to be concentrated like that. What we need actually is to ultimately end up with, and I said it was a metaphor and analogy. You want to end up with having all of the different variants of vehicle you can possibly imagine that fills out all of the utility curve

that exists in this space. And we need a much more adaptive, organic type of system than you're going to get from something that's command, control, and centralized and so forth. And you need, I think, what you can get from all of the great works being done within the Web3 community. So I think technically, it could be much better than the centralized alternative. However, if we don't find a way in which we make it economically better, and not just economically better in the long term, but in the short term, for small groups of people, you know, to your point, Sebastian, it might be hard to win because the kind of follow the money forces are very strong. So that kind of covers a lot of the decentralized ownership aspects.

We've also seen projects, and we ourselves at NOSOS has not been completely innocent in this, at actually putting agents on chain, right? So in the beginning we did things like, we said, okay, this is a prediction market, I'll give you 10 die. This is how it works straight on it, sort of thing. But actually we've progressed to the point where we can just, you know, spin up an agent on the chain and say, here you have 100x die, make me some money.

And the bot goes out and does just that. So obviously there are very real concerns here as to how wise that is, right? Do you want to put something that in principle could compete with a huge, and have different set of things it optimizes for on a chain that you can't turn off. Do you have thoughts as to that, Richard? I mean, I think, look, I don't fall into the camp of there suddenly being an acceleration towards the singularity when these things start teaching themselves. and that happens overnight.

That being said, the ability to teach itself an agent or a model to refine itself, you know, that kind of second derivative, that is something for us to watch out for. You know, I think Mustafa Salomon, for example, would be kind of in agreement on that. But, you know, I will leave those deeper elements of the kind of the precariousness of the,

kind of the AI gift to those who are focusing, you know, fully on it. I mean, but I will say, I guess, where it comes up is, do we want, you know, closed AI, let's call it that, owned by corporations and governments, where the choices made possibly by a very few people who has that power and who can harness that power? Or do we want something that is more openness, open and inspectable? I think I go for the open and inspectable and widely distributed option, and that maybe we even reach a point where we are in a dynamic equilibrium of mad, as they call it, of mutually assured destruction with these things.

Maybe that's some kind of end point we reach. But look, you know, knock on words, it seems to be in something that has actually, in some respects, at least helped out in the context of, you know, our transition. from the kind of wartime, second war, Cold War and the sort of nuclear destruction again, I'll touch with a couple more times. But I guess on the question of, you know, do we want to give effectively kind of persistent autonomy to agents on chain that can do things? It feels like it might be fine, indeed, unless it can teach itself to adapt into something

that you don't expect. And maybe the problem is we don't know. in advance whether that's going to happen. I think it's going to be quite hard to predict what's going to happen with markets. I think they're going to become increasingly perfect, but maybe they're going to become increasingly volatile as well. And then I think that the other side of the impact of AI on Web 3, just a very specific one, is that if we look at the way in which AI co-pilots are tracking,

and their ability to create applications, that we're going to go from a situation where there are relatively few solidity programmers and it's incredibly hard to audit smart contracts to a world where actually, you know, and we need to watch out for this, that vulnerabilities are possible to spot using AI, but obviously therefore we can use it to audit them.

And indeed it's possible to use multimodal developer co-pilots to create, you know, the complete stack of front-end, decentralized front-ends, middleware and smart contracts, poor, you know, whatever, you know, might be, some kind of, whether it's a gig marketplace or, you know, whether it's a decentralized exchange of some description. And therefore, the kind of community strength that some of these L-1s have sort of starts to fade into the background. So I don't know what you've been thinking about on that particular.

particular topic, Frederica, because you seem to be nodding your head a little bit, that it could change the dynamics, the competitive dynamics, quite quickly. Yeah, no, I completely agree. So I totally see the dangers. I think stopping is not a great option because we can't force everything, we can't force every participant to kind of stop, right? So kind of, I'd rather we with kind of like, hopefully good intentions, are in the mix as well, because otherwise if you say, oh, this is a little bit too dangerous for me, if someone else continues, that does you no good, you know, in the larger sense. Absolutely.

Certainly there are people I respect on the AI side, including Nathan Benech at Air Street Capital and, you know, folks, you know, who have commented that obviously the hewn cry around, you know, let's slow down on AI, let's regulate it. it is a kind of classic form of regulatory capture by, you know, big tech. And, you know, it's self-serving to say slow down if what you're doing is, meanwhile, you know, having a hearty breakfast and getting ready to sprint into the lead. You know, so I think, yeah, I would agree. I think we need to keep going and as many people need to keep going as possible.

And it's ultimately going to be extremely hard to put the connection back in the bottle. But I do think a lot of the principles around transparent, you know, governance, decentralized governance, incorruptibility of various, you know, points of governance, collective ownership, modern mutualism, if you will, in terms of we like, a good principles, you know. And in some, from Web 3 to apply to AI. And like, in some senses, you know, you might think that it's hard. to imagine of a bigger use case that in reinventing money and the world's financial system, but actually sort of reinventing, you know, reinventing, you know, and I think, you know, crypto still has this opportunity to reinvent our relationship with our personal money, but actually what's going to happen with AI is we're going to reinvent our relationship

with the use of our personal data. And I think the freedom to, you know, to have that personal relationship, their direct relationship is like effectively a human right. But not just a human right in a kind of, it's nice to have, but essential for the kind of balance we want in society. Yeah, I agree. I think the next 100 years and maybe even less, you know, we'll see like tremendous societal shifts caused by AI. And, and we need to really think as a society how the data is used and also like how the AI

is kept safe sort of aligned with our objective as a species and doesn't cost chaos which is a tall order. But yeah, maybe you're shifting gears here a little bit. What other interesting trends are you seeing in this space? And I'd like to ask you a little bit about DFI, because I mean there's been tons of innovation in defy

over the last four, five, six years. But it appears as though innovation is slowing down a little bit. There's not very much that's sort of new. It's like a lot of rehashed ideas and optimizations. Do you see defy, have we reached the top? Or is there more innovation to come by way of like new types of applications that really redefine our interaction with finance? So I think we're nowhere near the top.

I think, but I think that realizing that the capital that Defi can handle is not just your financial capital, but your social capital and different, and that, you know, decentralized applications are going to be intrinsically social and they're going to have the ability to also handle transactions within them and how those all get intermingled, I think is going to be the opportunity. I mean, we see people looking at building decentralized exchanges that can handle not just fungible also, but various combinations of non-fundable token.

And so I think that there is opportunity in that direction. and I think you'll see a collision between, should we say, AI's hunger for data, decentralized social media in a way it gives you the ability to, you know, mint data that you own, and defy's ability to create marketplaces. And I think there's an intersection there that's very valuable. And also, I guess, if you think about it, it's just not just not just content data. It's also things like health data. So if, and we've seen how DFI can, you know, and the world casino that we sort of once the Web3 has sort of stabbled into can be incredibly sort of good at creating these highly liquid marketplaces.

even if it is with money that is quite hot, as in it's there while the going is good and then it sort of disappears. And I think sort of one of the next challenges is to how to apply that to data sets that are actually where the kind of value is less, you know, volatile. But, you know, there's, for example, sort of health data. It's long lasting. But you can use the principles of creating liquidity and bootstrapping marketplaces that we've learned how to do on the defy side and have them kind of permute across. So I think in a sense, defy hopefully will become useful by being applied as a primitive, as a building block to the applications and marketplaces we're trying to build. And then in that way, we'll see it become much bigger. And I also think on its own

legs, I mean, like, I guess we could ask ourselves a question, you know, do we yet have an app? And I think, by the way, as a sort of sidebar, we probably feel that apps are disappearing. But do we have you know, a sidekick, you know, set of experiences, continuous sort of streaming application experience that is sort of tuned into the value that we have and all of the data that represents our credit score and that continuously surfaces for us from, you know, a financial marketplace where there is, you know, not human intervention and presents it to us for whatever opportunity we're looking at,

at that particular point in time. Well, that's not happened yet. And that seems to me the core, still the core open finance or decentralized finance opportunity. And a lot of the ability to do that is only just coming to pass. You're talking about sort of this emerging use case of digital asset managers that make use of AI agents to, find like sort of optimized positions for one's portfolio that that kind of use case.

Yeah, it could be optimizing the positions in your portfolio, but it could just be saying, you know, like I'm going to buy some clothes for delivery tomorrow. I'm going to, I need to use credit for that. Maybe I have free credit. Maybe it's only two days. If I've sort of said, I'm going to return it. But to negotiate the credit for you, you know, automatically. in the background as part of that transaction. Also simultaneously, of course, like capture a tokenized set of, you know, information about that interaction

that becomes attached. It's part of your kind of data resource that becomes attached to your profile, which then in turn feeds into a future possible offers of credit that you will be given sort of instantaneously next time you need to do that. Or indeed, not just credit, incentives, loyalty, whatever. I think that the loyalty space, which was something I remember looking at a company

called Offermatic, best part of 15 years ago, which was looking to kind of use your transaction data across all of your different accounts to give you better, you know, loyalty. You know, surely we want to get to the point where all of the information, you know, around who you are and how you transact is used productively for you, hopefully, to, you know, to give you the offers and the loyalty and so forth that you want. And then on a kind of permissioned basis, you get those offers that you really, you care about rather than having a ad revenue driven model that's sort of inbound. It's more kind of permissioned marketing that takes place.

So I think all of those are, you know, open commerce, open loyalty, open payments, you know, and the defy marketplaces behind them, they're all interwoven. and I think we have barely got anywhere with respect to those so far. And within that is the kind of real world asset marrying to any concept that exists today with that world is tricky.

And the reason why obviously it's easier in the kind of wealth casino cases when you have a fungible asset and it's all natively on chain and you're staying on chain, that's the easiest place to get those marketplaces going. But I'm sure we will get there. Yeah. So I'm speaking about trends here, maybe looking back in your

journey as an investor in the space, which trends, which notable trends did you see over the years that you thought would play out like you thought would become a larger part of the industry, but ended up not? And, you know, what did you learn from, uh, from those. Yeah, I mean, I think, I just mentioned one, which is, I think the loyalty was one that we thought would, there was a, a put of payments and loyalty company that we incubated that was Web 2 called Yo-Yo. I remember trying to pitch my partner, Al-A Feliz, to take that on chain back in, I don't

know, 2014 or 15 or something. and you know none of that has taken off as quickly as it might again you know I think it's really a question of time or timing rather than anything else I think that on-chain games is another one that that obviously I think has got challenges to take off and a lot of people know we're invested in cartridge a lot of people are doing a lot of work other firms like lattice and so forth to to build the capabilities to make that work and have sort of AAA standard games that that um they can operate scalably kind of you know on chain and indeed to to kind of allow people to own not just sort of buy um the various artifacts in those

games and then build kind of sustainable token economies rather i think yeah i think the jury still out on gaming it's uh yeah exactly it's not it's not it's not it's not it's not happened so i think that's been on uh so i guess we've been impatient for that that stake off i think um So I've always been bullish also on just generally the power of token economics, which I mentioned earlier, I think is an essential part of how we try to overt sort of massive centralization or acute centralization of kind of AI systems. And I think I do think that there's still a lot of work to be done to work out how to build resilient, sustainable token economic models and to value them. through different inner stages, because I do think it's fundamentally different from, you know, discounted cash flow, quite linear economics of a centralized company versus the self-reinforcing effects you get in something that is more akin to a city or a forest and it's got a lot of all these

different immersion properties and that arguably, or at least we would argue, can be, rather than inevitably going from kind of productive to extractive, as is, you know, much talked about in our kind of sphere. can be, you know, when we get that kind of platform power and platform risk, can be more positive some game and more sustainable long term. I think that's an area that we'd, I guess I expected us to advance a little bit, you know, further into today, but I think it's still still to come. I mean, we were back as of, you know, Sir Rare back in the day.

That was one of the ideas that was kind of elegantly simple and sort of was almost a kind of a trite example of what would be powerful on a kind of blockchain and has gone well. Shout out to the strata mafia. There you go. And I think one that we missed, I guess this was a little bit to do with, a little bit to do with actually trying to apply a more typical venture capital approach to the space was that, you know, in venture capital in general, in the sort of famous case study of this when was it when a 16c backed bourbon that became Instagram and also pick a pleas if I

remember what it was called you know and then they ended up on a kind of collision path and obviously Instagram was kind of the the winner um you know that the vCs can only um can only back one player you have to pick a you know uh some to partner where then you're really kind of full on with that whereas i think in the kind of obviously l1s and all of the space a lot of a lot of folks have backed multiple different players and then there's been enormous the opposite direction there's been enormous value generated and possible returns distributions for venture investors from a whole swathe of those l-1s and l-2s so that's been a bit you know i don't think everybody predicted that that would happen that it wouldn't all kind of become an ethereum game quickly or you know or some

new you know winner and and so forth so i think that that's been not i don't think anybody call, well, we could chat to someone who said that they called that. But we, but we, we didn't, you know, play right across the field like that. So there was also kind of, should we say, a missed opportunity in that sense. Are there any specific deards where basically the company pitched you and you passed and wish you didn't? I mean, look, there were definitely, I mean, At the top of my head, there are two companies that I, by the way, I'm very bad at the media recall of these things. But the two ones that are on the top of my head is that, you know, we have a lot of time

for the team at Jensen. And when we first saw them, they were more enterprisy and focused, but we became decentralized and more, you know, we think they're doing great things. And I guess the bit I'd like to emphasize that, you know, and talking to, to them, to the founders, Ben, and Co.

that they, um, they're just very impressive founders and individuals. And that's your huge thing that we can have, you know, index on. And you've got to be very careful to, to,

um, to not talk yourself out of some of the challenges in these companies. And then another, another one, um, that we should have dug into more, um, probably is,

is lazy ledger that became Celeste, um, uh, where Mustafa actually, um, We had a discussion about him joining us very early when he was still at UCL. So we knew him and we knew other people in the space, but that didn't happen. And look, we're, as we've already been discussing, like, kind of, should we say DPN and DPN for AI.

And, you know, we think that's an important part of the future and interoperable modular open blockchains. We think that's also an important part of the future. So we definitely kind of, there was on pieces. Well, I'm going to I'm going to now spend the next few minutes that we have here asking my own selfish questions as an emerging manager of a VC fund. So yeah, the first thing I wanted to ask is like, you know, crypto, crypto valuations obviously have a premium. And that's because of the low time to liquidity. I mean, that's one of the reasons. at the same time, you know, VCs present themselves to teams as long-term aligned and like long-term

supportive. Do you think that there's a contradiction there where, you know, short-term liquidity and the way market cycles operate and it forces you, if you're, if you're, hence, it is as a fund to return cash to your investors, return capital, it is, is the, is the, is the, is the, is the, short-term horizon of liquidity in crypto counter to what a fund should, in theory, be providing to a team is also like long-term support. Yeah, how do you think about that? And specifically when it comes to exiting liquidity. Yeah, so it's definitely something we've been wrestling with six, for six, seven, eight, eight years. And obviously, look, it's a new frontier,

and so therefore there's no reason why there should be a, a well understood or, you know, a quick or easy kind of answer. And at risk of kind of giving a kind of, it depends, or it's somewhere in the middle answer. I think that venture firms should be, you know, venture backers should be thinking for the long term and backing those projects for the long term, you know, I think there's clearly a segment for which that makes to no total sense.

I think, and that makes sense for the projects because they want to know that you've got their backs through thick and thin. You know, there have been a, you know, when it comes to the question of when to sell, you know, you also have a duty obviously to your limited partners to try and deliver returns. And when you come to raise your next fund, you're going to be asked where your distributions are. So there's a pressure on yourself to be able to achieve that. That's definitely easier said in a sense than to do for a few different reasons. Well, number one, that tension with the long-term interest of the project. So I think what we have tried to do is that you need to have that relationship with the project

to understand how they are hedging. And in some cases, it may well be the case that they need to hedge. They should be said to be hedging their own treasury in order to, provide for the solid basis for the long term sustainability of the project and certainly it's fair if there is that hedging going on

for you to perhaps be similarly hedging your position at that point in time but that obviously involves a dialogue and historically maybe that dialogue has not been not been there I think it's in the future but it is a difficult dynamic to manage because projects may not want to talk about

the fact that they are hedging because that itself can have it in part packed on pricing. I mean, there's another dynamic for venture investors, which is that it can be said, and it has been said, that it's actually easier to buy than it is to sell. So, you know, so backing a great project and getting in is one thing, but knowing when to decide that it's time to get off the train, it can be very hard as well because you think there's more to go.

And then afterwards when the market's coming down, you can feel like you're trying to catch a pulling knife and whatever you hesitate. So I think that muscle needs to be exercised. There needs to be discipline around that. So it's a different reason in which it can be difficult to get it done. So you need to be sympathetic to the founder of the project, be integral with it.

And then you need to actually have the discipline internally, the processes, the people, the numbers, the dashboard, whatever to think about it. and to structure it. And I think it probably comes back to the fact that expectations are always key. If you can be as a backer, you could say, look, I'm with you through thick and thin. But if there is a point at which the market, you know, the liquidity has occurred and you've done extremely well with the kind of token price, then it might make sense for us to hedge our position. And this is how we're going to do it. And we'll make that collaborative and so forth.

So I think communication and expectations would be crucial for making sure you remain aligned with the founding team. Yeah, that's helpful. I think essentially requires a measured approach and also proximity and communication with the founder. That's helpful. And one other thing here, like, you know, you guys have to look at so many deals over the years, like possibly thousands of deals invested in more than 100. What are the criteria that you look for? So like when a deal comes across your desk, what are the main like three things that you look for

that, you know, will, we'll jump out at you as like, this is a potential 100X? Is it the team or is it weighed more towards the tech or like really true product innovation? What are the main things that will jump out and, you know, cause you to say like, okay, this is a priority. like we should pursue this as possibly like, you know, going forward with us. Yeah.

I mean, so clearly this is a topic that comes up, you know, with many VC-oriented podcast or conversations. And for me, at least, if you're talking about at this stage investment, at the end of the day, it's, you know, just the team, the team, the team. I mean, it isn't because you need to see some evidence. of how they're thinking about their initial product offering, is this a clean, sharp insertion point that's going to get immediate traction and give them learning

and is that elegant in the way they're constructed it? And you need to see evidence that they're thought about, maybe not the initial market, but how they can tack to creating even a phenomenally sized to the market over time. But in a sense, those are just evidence points that the team you're talking to are just incredible. They're not, it's not necessarily the answer, you know, the specific answer.

It's the, it's what's going into generating that answer and the fact that that indicates that they'll be able to come up with the next right answer when new data represents itself and they work out what to do. And so, for me, that's the kind of crucial thing. I think one of the things to kind of, to kind of, to, to, to, to, to, to, to, to, to get your head around as early stage investor. And my friend Fred Destar, I think he called his blog post, he runs stride, you know, talking about how I learned to get comfortable with risk and embrace it or something like that.

You know, and it is about getting comfortable with how much risk there is. And in fact, particularly, of course, if you're building a venture portfolio, we are actually looking for investments, for projects that are taking more risk than is reasonable for the founding team to be taking on a standard. to learn basis because we we want every every single time that project to be really swinging to be totally extraordinary in the outcome that they produce but it's not just risk it's also a kind of form of messiness that that I think can exist so and I don't actually thinking about it you both have these teams that are incredibly capable and and have thought things through and can evidence that but at the same time getting comfortable

that there's great risk, uncertainty, and even messiness at the early stage of those outfits. So if you, if you, I'm answering the question of what you're maybe not looking for. If you're looking for everything to be buttoned up, then, and if you're looking to ensure that you can then feel comfortable that the things at the end of the day aren't going to go too wrong, that's the wrong thing to be looking for. We're looking at that if a whole series, if the sun, the moon, the stars and the, what, whatever, you know, some other celestial body kind of line up, then how, if that all happened, how incredible it could then become, that's what you're really attuned to.

But back, actually makes me think of another point on, actually quickly on this element of liquidity, which is different and distinct for, for the few capital allocators who are out there, you know, putting lots of money into venture at the moment in the world, let alone Europe because we still do have a bit of a dearth of that, looking forward to some interest rates red cuss. But the other thing is that we've observed is that because of this kind of organic, more city-like, positive sum set up of these networks that get built, where to take the canonical example, we don't know who Satoshi is,

we don't know who the founder is, you don't have some key risks that you might have in normal startup projects. And so that actually the fallout rate, the kind of fallout, the kind of go-to-zero rate is lower than you would see statistically from normal kind of startups because of that kind of open, collaborative community-based nature, I would posit of these projects. And so, yeah, so you've got the earlier liquidity,

you've got the lower fallout rate, but you still, nonetheless, if you really want to deliver exceptional fund level returns, need to be swinging for the fences as the proverbial phrase goes. Well, thank you so much for that. That's great insights, and this has been a really fantastic conversation. So, Richard, thanks so much for finally coming on the podcast. Hopefully we can get you on again in less than 10 years.

And, yeah, we'll catch you soon. Well, thank you. I plan to still be building fabric in 10 years, but let's not leave it 10 years. Thank you. Thanks, Richard. Thanks, guys. Really enjoyed it.

Thank you very much. Thank you for joining us on this week's episode. We release new episodes every week. You can find and subscribe to the show on iTunes, Spotify, YouTube, SoundCloud, or wherever you listen to podcasts. And if you have a Google Home or Alexa device, you can tell it to listen to the latest episode of the Epicenter podcast. Go to epicenter.tv slash subscribe for a full list of places where you can watch and listen.

And while you're there, be sure to sign. that for the newsletter, so you get new episodes in your inbox as they're released. If you want to interact with us, guests, or other podcast listeners, you can follow us on Twitter. And please leave us a review on iTunes. It helps people find the show, and we're always happy to read them. So thanks so much, and we look forward to being back next week.