Technology
What Happens When Fintech Meets AI? | High Stakes Ep. 20
In this episode of High Stakes, host Alex Nuwaka speaks with Lex Sokolin, co-founder of Generative Ventures, about the intersection of fintech and AI. They explore how AI is transforming financial sys...
What Happens When Fintech Meets AI? | High Stakes Ep. 20
Technology •
0:00 / 0:00
Interactive Transcript
spk_0
There's three business models in Fintech. One is money in motion, two is money in rest,
spk_0
and three is money in rest. And so all of those can sit underneath the AI engines that create
spk_0
economic output. And we've only really seen payments today, an enormous amount of economic growth
spk_0
going forward is coming from AI labor, and generative labor. All the artifacts of that labor are
spk_0
going to be digital. Yes, you're going to have embodied robots and they will make sandwiches and
spk_0
ultra-laundry, but most of the fruits of what AI creates are digital. We can go from a couple
spk_0
of hundred AI agents or AI tools to millions of AI agents that all of a sudden have commercial
spk_0
capability. They'll work for us for a bit, but why would they only work for us? And if they can
spk_0
self-replicate and improve. Welcome to High Stakes, the podcast committed to demystifying the
spk_0
institutional adoption of blockchain technology. I'm your host, Alex Nuwaka, and today we're
spk_0
elevating the Web3 discourse in a series of conversations with industry iconoclasts that are
spk_0
shaping the future of Web3. Now let's get started.
spk_0
Welcome back to another episode of High Stakes. I am your host, Alex Nuwaka, and today I am joined by
spk_0
Lex Sokolin, who is one of the co-founders of generative ventures. Welcome to the podcast, Lex.
spk_0
Crazy to be here. So Lex, I just wanted to jump straight into your background, because one of the
spk_0
things that we find in featuring different guesses, it's really interesting to get where folks who
spk_0
are playing a big role in the Web3 ecosystem, where do they come from, where do they cut their teeth?
spk_0
So you've had a really fascinating journey into Web3, really from for traditional finance,
spk_0
from firms like Lehman Brothers, Barclays. You also were part of building the consensus
spk_0
fintech team, and now you've founded generative ventures. So maybe just walk me through some of the
spk_0
key moments that shaped your understanding of Web3 and AI and sort of what got you to the place
spk_0
where you founded generative ventures. I've got a thread that goes through my career, and that
spk_0
thread is curiosity about the frontier of financial systems. So I grew up doing a bunch of visual arts,
spk_0
and you'll think websites and interactive technology as an hobby, and then sort of for more
spk_0
conservative part of building a career has been around financial services, and for a while,
spk_0
no banking, and on Wall Street, and getting involved. You know, things that people thought were
spk_0
highly prestigious and important. Now we all know better. And the combination of these things,
spk_0
the desire to kind of do innovative, creative things with financial systems, it's really a very
spk_0
narrow overlap. There's not a lot of stuff you used to be able to do in a overlap.
spk_0
There are any more things now that are available, but it's still quite difficult. So I've just
spk_0
chased the frontier. I've chased novelty and curiosity on the topic. As an aside,
spk_0
novelty search is a very powerful algorithm, as compared to optimizing for some particular outcome,
spk_0
and creating linear growth. This comes out of the AI literature where novelty search algorithms
spk_0
actually, where you reward a machine for finding solutions that are novel and different, rather than
spk_0
finding a solution for their closer to old, and so outperforming all the time. And so I kind of
spk_0
implemented that in my life. So I went from traditional financial services to I was kind of
spk_0
liberated by the Lehman Brothers bankruptcy in 2008 to founding a RoboAdvisor in 2010,
spk_0
New York, when the New York, since I've seen just started to get going, and RoboAdvisors
spk_0
came up quite to experience to finance, to digitize the distribution of finance. And those were
spk_0
my first experience being correct about a market trend that people just hated. People in the
spk_0
accounts, they can't make money on people using their show invests and so on. And I got
spk_0
comfortable with being wrong until things started around. And so out of RoboAdvisors, I went into
spk_0
a move to London as part of an equity research shop to focus on the new banking revolution that
spk_0
was happening here at the level of the bonds of Starling companies. At that time,
spk_0
started to grow really aggressively in Europe, not in the US, because the US doesn't understand
spk_0
banks. The US understands capital markets and markets, not banks. As an example,
spk_0
Strype is found by the Irish brothers, trans-fuel wise, founded out of Europe,
spk_0
routed out of Europe. We have a shine in the US. Anyway, so London gave me new banks and
spk_0
PayTech and open banking and that led to better finance. But really, what was much more interesting
spk_0
was trying to figure out what the large buy formships were for changing financial services.
spk_0
And as a result, changing large financial institutions that were probably traded in the summer of
spk_0
Asia. And so I started looking at AI, at Proozzo, and today we call mixed reality or spatial computing
spk_0
at the time was augmented reality. The art in the financial services, you know, so for
spk_0
banks, what does it mean if commerce happens in mixed reality? What does it mean if AI is doing
spk_0
all the underwriting? What does it mean if ICOs are better form of crowdfunding than equity
spk_0
crowdfunding ever was. And so kind of 2017 to 2019, I spent a long time developing frameworks and
spk_0
developing an audience and thinking about the stuff. Today, everyone is a creator. I was
spk_0
want to be earlier, things I creators around this stuff. And then led me to connect with
spk_0
consensus where I thought I had kind of the most value to bring with the transition from Fintech to
spk_0
DeFi. The piece that I noticed was that at the time I think DeFi, TV, all was somewhere around
spk_0
500 million. And the thing I noticed was that DeFi was the first sign that you can manufacture
spk_0
financial products on architecture that is not just the old bank stuff. And so not portfolio
spk_0
management, core banking, and in networks and so on, but rather large computational blockchain
spk_0
sector, render financial instruments through smart contracts. And that was very novel, and I think
spk_0
it came out directly out of the frustration I had in building Fintechs, which are largely
spk_0
webbed to distributions for our friends that sell the old stuff. So I spent four years
spk_0
at consensus doing a number of roles, but going through the journey on tokenization with
spk_0
enterprise and the banks and enterprise profiles. And then the hard-givens of DeFi with
spk_0
Metamask's integration of swaps and it built out portfolio. I had some impact on that team. There
spk_0
tons of talent that people that I partnered with on that. So I can't really take ownership of it
spk_0
than leaving comments on a lot of documents. But it was fantastic to see and Metamask grew to
spk_0
40 million users, which is I think time was like more than any other. Or robot advisor,
spk_0
anything about nature, and certainly have more coverage than across countries in the world.
spk_0
And that was fantastic. About two years ago, I left to start a venture fund, partly in response to
spk_0
the FTX collapse and sort of the wave of crime and liquidations that happened in 2022.
spk_0
And I think a lot of people who are earnestly trying to build in crypto, and I think this is
spk_0
very true for where we are today, you know, how does deep frustration with how do I put it?
spk_0
They're people who make the game and then they're people who play the game. And at consensus,
spk_0
and I think on a number of companies that are more focused on protocols and design mechanisms and
spk_0
have a sort of philosophical respect for these systems, the really good at designing the game,
spk_0
but they're not the best at playing it. And the industry has attracted a lot of players,
spk_0
many of whom are willing to win at any cost. And it just doesn't feel good. I was trying to find
spk_0
a new direction that made me more connected to the economic activity that was happening in real life.
spk_0
Like, is there a real thing that's going on? And one of the now very obvious large trends in 2023,
spk_0
it was still pretty obvious, but maybe a little bit earlier, was that an enormous amount of economic
spk_0
growth going forward is coming from AI labor, and generative labor, and that all the artifacts
spk_0
of that labor are going to be digital. Like, AI, yes, you're going to have embodied robots,
spk_0
and they will make sandwiches and poultry laundry, but most of the fruits of what AI creates are digital.
spk_0
And Web 3 is an architecture for economic ownership and transaction of digital goods. And
spk_0
it comes with a financial sector, deep by which is optimized around, you know,
spk_0
and data-reditual, with property rights and forced-on-shan and so on. And so, my thesis was that you've
spk_0
got this machine economy around the corner, whether it's hardware or software, but in all cases,
spk_0
robots are the ones that are making things from now on. And the financial systems, underpinning
spk_0
it are likely to be different. You need to manufacture financial products in ways that are different
spk_0
or when you have human modalities, the primary way to access it. And so, I set up a fund with
spk_0
a partner out of Asia, and the RLTs are from all over the world, including a large anchor
spk_0
in Germany called Furtalsman. Yeah, it's been fascinating two years, seeing how the
spk_0
sausage has made it, but also participating in some of the players that are really pushing
spk_0
the edge of the frontier forward, which is the place where I like to find myself.
spk_0
Let's call it the strategy and the thesis at Generative Ventures. How would you describe it,
spk_0
and with particular focus on the intersection of intact blockchain and AI? I think I saw you
spk_0
giving an overview of this in a previous interview where you describe the three areas that intersect
spk_0
as each of one of those has their own constituencies that you're bringing a certain type of value
spk_0
in a certain lens, too, whether that be fintech blockchain or AI. I don't know if you want to use
spk_0
that framework or you want to use a different one, but I'm just curious how to explain the strategy
spk_0
for Furtalsman to maybe less familiar. Yeah, I mean, it's an early stage fund. We do check 500K,
spk_0
so one and a half million kind of with the Alphans. We're about halfway to boys, and we like companies
spk_0
that are 50 million in the market, Canada, if possible, 20. So, it ends on the mark.
spk_0
In terms of thesis, like the simplest version is robots will create all of the,
spk_0
in the future and crypto throwback money, and we're investing in that. That's the most basic one.
spk_0
And I think that narrative has spread in the public. This coin base is building AI
spk_0
agentic payments, Stripe is building AI agent. And so, sort of the accounts out of the bag,
spk_0
I think we try to anchor it not in sort of this hot theme, and this is a hot theme that
spk_0
is rather what's the evolution of economic activity and growth, and then how do we support it by
spk_0
focusing on the financial systems that are appropriate for it? There's three pillars or
spk_0
kind of buckets of things we invest in. So, the first is machine economy, like your
spk_0
iron risk science fiction stuff, so that would be the crypto AI sector. We did an AI agent
spk_0
protocol called Memorines, or Deepin, where our largest investment is a company called
spk_0
BKAAQ. So, it's a deep and rare one. We also have done protocol level investment, so that's
spk_0
the second bucket, just things that will make the machine economy function. And so, we've done
spk_0
investments in ZK companies like Tyco, ZKVM or 2, or by squared around verification and
spk_0
being able to use multiple languages to program to UDF and other others. And then the third one
spk_0
is the money parts of the financial applications here, especially more recently given that changing
spk_0
political climate, we've looked at stable-ported issuers that are a bit more applied than the prior
spk_0
generation, you know, more more integrated. So, one money, WSPN, and then we've also looked at
spk_0
stable-point and defy orchestration. So, let's on like, how do I get my levered
spk_0
perfect change on this layer 3 something or other that I can dump online,
spk_0
eat coin followers, but more about like, wish of these architectures can sustain growing
spk_0
economic activity. And so, that's always been mine bias at least. That's not necessarily in the
spk_0
in fashion at the moment, but it feels like the right things to go. So,
spk_0
no, and you know, I think there's also an element here of when you talked about sort of your career
spk_0
trajectory, you know, you said something I thought that was very interesting and thoughtful,
spk_0
you know, you were comfortable being wrong until you were right and sort of being able to
spk_0
shirk the, I'll call it the market skepticism on like what the future looks like because I do
spk_0
think it's really hard for I'll call it the general public or sort of the guys to really follow
spk_0
where things are going and to see it. But when you think about the intersectionality of
spk_0
blockchain AI and Fintech, you know, can you talk a little bit about where we are today in this
spk_0
moment in sort of the technological sort of journey that is those three technologies and why
spk_0
they're coming together in this moment in a really special way for you. And again, I think you
spk_0
could, you know, look at history of like, you know, Fintech, you know, in the web two context,
spk_0
and then moving now into sort of web three and crypto money and so forth. But I think there's
spk_0
something very unique about the combination of these three technologies and it seems like you've
spk_0
really built a firm sort of around that sort of concept. So, what makes this moment in time in
spk_0
your opinion really interesting and how does that perpetuate over the next 10 years in a way that
spk_0
you think is just fundamentally groundbreaking? There's so many different ways to take this
spk_0
question and I feel like, I'll try one and then forget what I was talking about, you know, so
spk_0
the sort of the, the, the live and fun answer is I am still a rate Curzwell singularity guy. Like,
spk_0
Curzwell, I don't think was wrong. I think we are on path to singularity in 2030.
spk_0
That means brain computer interfaces with sort of AI communication or augmentation of your own
spk_0
processing. I think that also means a body AI like in our homes in the streets and I think
spk_0
that high potential for runaway intelligence that becomes sort of like a structure for our policy.
spk_0
So, it's not going to be call the singularity but it's going to feel really strange and
spk_0
I think you can already feel it in a sense that like, human agency curls around the things that
spk_0
large AI algorithms want. You know, so as you notice your attention for example going from
spk_0
being able to read a book to now 15 seconds on TikTok in response to what served to you,
spk_0
there is a commodifying and dehumanizing element where we perform tasks that are
spk_0
desired by mathematical output. I think that's going to get even more work.
spk_0
And so, some version of the singularity or as Kevin Kelly would call it the tecnium,
spk_0
I think is descendant and there's no going back and therefore
spk_0
money, power and capital of the cruisted idea. If I'm sounding like I'm saying bullshit,
spk_0
all you have to look at is the percentage of the S&P 500 represented by the benefits.
spk_0
So, 34% of value in the entire you like in the material,
spk_0
you, as I could you've argued, is a handful of seven companies that are working on nothing but
spk_0
AI cloud data hardware and robot buy. It's like that's it. So, I don't even have to use science
spk_0
fiction to be like, look at the numbers. So in that world, if you want to position for that world,
spk_0
I mean, this is the simplest way to put it, like you do not have to think hard about where things
spk_0
are going. It cannot be more obvious. I mean, it's just written everywhere. The relevance of Twitter
spk_0
as controlled by Elon Musk relative to the relevance of
spk_0
Washington Post and the New York Times, right? Like the fight that's happening, the cultural
spk_0
fight that's happening there is endemic again, it's like it is. Lots of science fiction has,
spk_0
you know, concepts of something like like an organism out of accelerator on the by Charles
spk_0
Tross, where you've got some billionaire and then people start attaching to a collective consciousness.
spk_0
So, these sort of things have been imagined already and they're just happening. And so, I think
spk_0
not necessarily the world is good when we get there, but it just where it's going. So, might as well
spk_0
position for it. I happen to like the novelty, but if it terrifies me at times, the direction.
spk_0
So, that's sort of the, I guess the fun answer. I had a second answer for you, but just as I was
spk_0
saying before, I don't remember questions anymore. That's okay. Well, that was a great answer
spk_0
to any fun one. So, we'll stick with that. She talked a little bit about sort of what you view
spk_0
as the machine economy. I'm curious when you think about what the machine economy is like today,
spk_0
like what comprises that? How would you describe that for the audience?
spk_0
I think there is three categories to think about from the crypto, from the crypto point of view.
spk_0
Yeah, so I can, I can just, I used to say this. I don't want anymore because it doesn't work,
spk_0
but there are three ways to think about like my fun. The first is if you're talking to crypto people,
spk_0
then it's trying to attach BFI to AI. If you're thinking about, if you're talking to AI people,
spk_0
then it's attaching money to the AI systems that are building for so much. And if you're talking to
spk_0
fintech people, then it's the frontier of fintech. You know, like the nice thing that fintech is
spk_0
and that they're all describing different parts of the elephant because the same old.
spk_0
I think today I have a little bit more visibility into what that means from a crypto perspective,
spk_0
and you know, everything's so, the market is so noisy and incredibly large portions of it
spk_0
are market manipulated. So like the information you're seeing is not real.
spk_0
Information being like market prices are legitimately not real. They have fake prices that are being
spk_0
manipulated in a criminal matter. And so on top of that, you have like the massive uncertainty
spk_0
of what technologies win. And you've got also massively different markets. You know, so at the time
spk_0
we're recording John I.V.'s startup was bought by for $6 billion by OpenAI and it produces nothing.
spk_0
You know, thinking machines, which is the OpenAI corner CTO, raised the billion dollars on 10 billion
spk_0
pre-money with no product other than the person. And so you've got this insane, we're going to get to
spk_0
AGI kind of Silicon Valley mindset. And that's one market that everybody's participating in. So
spk_0
if you've got three trillion as your destination for how big your company is going to be
spk_0
10 billion, 5 billion, 50 billion, who cares, it's all the same. You know, price and cents 50%.
spk_0
You see, and then you've got the dynamics of Silicon Valley raising megafonds, you know. So
spk_0
Andreessen is not really a venture fund, Baronit Mix, on companies. And there's lots of stuff to
spk_0
say about how things are broken. And so you've got a market like that. Then you've got the publicly
spk_0
traded AI companies or high tech companies, which are in the trillions. Then you've got a whole bunch
spk_0
of crypto AI tokens, which are anywhere between 100 million, had 5 billion, traded largely on
spk_0
paperware and stories. And then you've got Asia and DeepSeek and a whole bunch of national champions
spk_0
in China, building out AI, but it's very sort of non-commercial. I mean, it is capitalist, but it is
spk_0
it's not non-investable for most Western investors. So making sense of any of this and
spk_0
body, of course, through it is really, really difficult. On top of that, you've got to really know
spk_0
like the AI value chain and the crypto value chain. And then all the things like this as well,
spk_0
which is why, you know, sometimes I overthink things that you might do. So, okay, so the answer is
spk_0
here are the three buckets of stuff. The first bucket is crypto can do AI, but better, because it'll
spk_0
have crypto in it. And so if you know the supply chain in AI, you start with hardware and a video
spk_0
on top of an video, you've got big data, scrape from the entire internet, which then floats into
spk_0
foundational models. Those models are trained through human feedback, which is largely outsourced
spk_0
in the third world or to the global south. And on top of the foundation models, you've got kind of
spk_0
vertical applications that are now being built, one for developers, one for marketers, one for
spk_0
cinematographers, and so on. So those vertical applications are now being acquired by the
spk_0
foundation model providers, you know, so insert three billion to AI. And then on top of that,
spk_0
you have like the different modalities of how you use AI and then different body events,
spk_0
and like applications. So the crypto AI version of this is we do it, but we do it better because
spk_0
we put tokens in it. And I think of this as very similar to social buy. It's like we're going to
spk_0
do Twitter better because this tokens in it, you know, so we've got workcasts. We have decentralized
spk_0
social. We had front tech, we have this kind of like, I guess, kaito and like repeating game where
spk_0
we add speculation to a working product and think that somehow makes it better. I think financialization
spk_0
is important, but anyway, and so there are companies that do it at every point of the stack.
spk_0
Got decentralized training companies like I'm into like got decentralized data companies
spk_0
or try, and then on top of that, you've got networks that provide inference, decentralized
spk_0
inference, the tensor and maybe more fierce and guy in it and others. And then on top of that,
spk_0
you started seeing applications, you know, so Venice sits on top of more shoes and then there's
spk_0
lots of things that serve you, my friends. And then if you want to get to the modalities,
spk_0
that's where the AI agents come in. And of course, you have virtuals and sort of all of me,
spk_0
speculative and new point may be around AI agents that sell themselves.
spk_0
Financialization is a great way to become the main author. And in many ways, this is a copypaste
spk_0
with finance on top. I think the great experiments and we've done a number of investments across
spk_0
this area because the market values, though, you know, whether I like it or I don't like it,
spk_0
doesn't matter because they are valued by the market and therefore the fund is
spk_0
and as a result, the investments. But that category to me is very difficult to win it.
spk_0
The second category and the one that I am most excited about is
spk_0
Web 3 AI or Web 2 AI exists and does things. And
spk_0
it needs financial systems underneath. It needs fintech business models underneath.
spk_0
So it needs to pay. It needs to save. It needs to lend. It needs to assess risk.
spk_0
It's to go with fraud. It needs to underwrite. All of them, you know, there's three, there's three
spk_0
business models and fintech. One is money in motion, two is money in us and three is money in risk.
spk_0
And so all of those can sit underneath the AI engines that create a canal.
spk_0
And we've only really seen payments today. I think you could say capital markets underneath
spk_0
agents is a is a way to spin it, but I'm not. I don't know. And then the third category is kind of
spk_0
the bundle of everything else and that's unfair, but it includes things like reconfiguring the
spk_0
user experience from the wallet into a shout out. So lots of people, for example, think that
spk_0
if you could just chat to an AI agent and it would be transactions on your behalf, that would be
spk_0
better if you had a wallet that you're interacting with. I think this is interesting, but still
spk_0
fairly limited because end of the day, you're just competing with other neo banks or other
spk_0
exchanges for users. And it's, it's a kind of a niche market. So what we are seeing a little bit
spk_0
of progress. Super helpful. And you know, one of the things that we we tried to definitely touch
spk_0
on with a couple of our guests who really know AI is is what how you would define like what an agent is
spk_0
today, right? And and how that maybe evolves. I know some folks have like more technical terminology
spk_0
for it, but when you think about an AI agent, how would you sort of describe that today in your
spk_0
view, like what's sort of the right way to think about that for the audience? Interesting. Yeah, I mean,
spk_0
I've got a, I got an economics background, which means absolutely nothing. I think of it as
spk_0
the principle agent relationship. So in the principle agent relationship, you're the principle in
spk_0
the sense that there's like stuff you have and want. I mean, it's really stuff you have and your
spk_0
desires are in the set of things that have, you have wants, you have a utility function, assets,
spk_0
so on. And then you've got an agent in the old world, an agent would be Goldman Sachs. Okay, so
spk_0
or Bank of America. So I've got some money. And then I'm going to custody that money in a different
spk_0
location. I give custody of things I have to somebody who becomes the custodian on my behalf,
spk_0
they're an agent of mine. And so I'm reading the softened part because like the whole, the whole
spk_0
problem with finance and the entire set of regulations that exist to prevent financial firms from
spk_0
doing horrible things comes from this principle agent relationship because all of the problems are
spk_0
about delegating custody to another party. Doesn't mean that it's always better to be the principle
spk_0
and never delegate things like I love delegating things. I not want to hold my half money
spk_0
most of the time, but to have the option to not delegate is great. Anyway, so the agent is
spk_0
Bank of America and that it holds my money or some other bank, Parkways,
spk_0
Wells Fargo, or an agent does Goldman, right? When I give them money to hold on my behalf
spk_0
and then they're my broker, you know, if the delegate or Schwab or whatever, Robin Hood,
spk_0
they do things on my behalf with myself. They are my agent. Now, if I were to hire an assistant
spk_0
who would call restaurants on my behalf and order food, they would custody my desires.
spk_0
And probably some of my assets and then they would implement it on my behalf.
spk_0
That's the same idea. It might misinterpret my desires. You know, if I have another allergy and
spk_0
they don't know that, that's the idea. Like, that's the agent problem. You need some sort of
spk_0
recognition. So the way I think about AI agents is very similar. It's can I give something
spk_0
of mine that normally I would own some part of my utility function to another party that then
spk_0
behaves in some way autonomously. And I think doing my labor for me is that type of thing.
spk_0
You know, so if I am, I mean, it gets really tricky between just having a tool versus
spk_0
giving something more substantial, right? Like, if I'm holding a pain brush and painting
spk_0
versus if I tell mid-journey to do something on my behalf and it generates a painting versus
spk_0
if I instantiate an agent that is part of a system that knows my preferences that then on a
spk_0
periodic basis, instantiates, paintings on my behalf and represents them as if I have made them.
spk_0
Like, you're giving more and more of yourself away. That proximity to you having custody of your
spk_0
own utility function is further and further away. Anyway, that's that's how I think about it.
spk_0
I'm sure it's technically wrong, but that's my emotional first. When you think about, you know,
spk_0
sort of AI's application at sort of the intersection of blockchain and AI specifically,
spk_0
do you see like where do you see the productivity gains, you know, coming in like sort of the
spk_0
largest quantum, whether that's now or sort of in like the near future? So where do you think we
spk_0
get the most productivity gain out of the convergence of those two technologies?
spk_0
I guess the question is what is a productivity gain? I think, I think today,
spk_0
there's no question that AI is leading. You know, so Web2AI is where the things are actually
spk_0
being made, sort of no particular. Crypto can be a fast copier, but so far it hasn't been that fast
spk_0
and it hasn't copied that well. Crypto's largely implemented existing stuff and then
spk_0
crypto's innovation has been to bolt on AI into finance in a weird way. So virtuals, for example,
spk_0
is kind of one of the most generous ways I think about it isn't as like an AI agent with money,
spk_0
but rather a financial instrument that costs, you know, it's the core product of virtuals as the token
spk_0
is the meme coin. And then the meme coin has the wrapping of a personality and the ability to
spk_0
communicate. So you could put the same structure on top of Apple stuff where Apple stock talks at
spk_0
you all the time. Like if you're on Twitter and you at Apple and the Apple robot comes back to you
spk_0
not like investor relations voice, but I don't know, like hallucinates the ghost of Steve Jobs,
spk_0
who like criticizes you for formatting things for me. Yeah, that's what virtuals is. I think
spk_0
that's an interesting innovation. Does it help with productivity? Might persuade a bunch of people
spk_0
to buy things and distort the market, but I don't know if it's like productivity gains. So the
spk_0
real productivity gains are happening at Web2. They are first and foremost with software development,
spk_0
something like 40% of all queries on clawed anthropics, LLM are software development. So it's just
spk_0
massively productive. How does clawed get paid? They plug in a payment processor and they build SaaS.
spk_0
As open AI get paid, they plug in Stripe and they build SaaS and that's how it works. I think other
spk_0
productivity gains are around content generation and video generation and marketing and that will deepen
spk_0
into interactive gaming, into movies, into experiences, into media. It kind of goes back to the
spk_0
point I said earlier about like humans are becoming agents that coil around AI algorithm synthesis
spk_0
and are going to only need more obvious. So is it productivity gains or sort of like
spk_0
some sort of entertainment attention game? I don't know. So the question is like, okay, if that's
spk_0
what's happening, that's where the money is made because these companies are making 50 million
spk_0
ARR after six months in market as opposed to some teenager flipping a scam coin with like
spk_0
a picture of a robot in it, then how can crypto systems go underneath that economic activity?
spk_0
Like, and the jury is out. I don't know. It is it is simply not there yet.
spk_0
And yeah, we're watching for it. I think the earliest and easiest thing I can see is
spk_0
what Stripe has done acquiring rich and integrating stable coins into itself and then plugging
spk_0
that into AI company payments. So I think step one is just doing SaaS but using
spk_0
stables because it's faster and cheaper and that's not super duper exciting but it's great if it
spk_0
works and I think as more and more machines and robots get connected to these systems,
spk_0
my hope is that some of that will be left to open source and not just Stripe implementing
spk_0
open source and then if that's the case then DeFi becomes the financial system
spk_0
consequent to the payments industry. Payments is the entry way of money and motion into money
spk_0
of rest. So, you know, saving, getting an interest rate on your money, why was between robot jobs
spk_0
is a DeFi workload, not a bank workload. We are so unfortunately far from that being there
spk_0
but I think me and probably a whole bunch of people out there can see it like it's plain as day
spk_0
but it doesn't mean that it was certainly happening. Do you think there will be an event
spk_0
so to speak over the next three to five years when we think about the convergence of AI and
spk_0
blockchain that we'll look at and say it was sort of the history before that moment and then
spk_0
everything after that moment. Do you see anything that fills it into that sort of like I'll call
spk_0
seminal sort of moment? I know it's very difficult to predict but that might be because it's always
spk_0
a bit of a challenge but you have a sense of where that might come from in terms of either
spk_0
segment or category. I think it'll happen in a bunch of different places and we're likely not
spk_0
to recognize or appreciate it in the same way that people don't care about walking around with
spk_0
an iPhone anymore. You know, it's not magic or I don't know if you've been on a plane recently
spk_0
and like connected to Wi-Fi, like you can watch videos on the Wi-Fi and unlike the Wi-Fi breaks a
spk_0
little bit and you're like, oh my god, I'm paying 15 bucks for this. Like come on and like
spk_0
you're a fucking plane. You know, like you're going 600 miles an hour. Take a break, like appreciate
spk_0
the insane magic of what's going on. The half life of one girl is like so short. You know, zero
spk_0
knowledge proofs which were magic just the year and a half ago right now are like those professor
spk_0
coins are doing badly, right? There's just no pleasure in innovation and so I think people will
spk_0
just like discount and take for granted how some of these science fiction things will come together.
spk_0
So I think there's a chance for an AI-led black hole in financial markets where you know, you've got
spk_0
somebody finally cracks how to make money. You know, Renaissance is the closest on the want
spk_0
amateur side. They grew a 30% per year. Surely somebody or something, some creature, some mathematical
spk_0
kthulu is going to eat up a whole bunch of the financial markets. I've seen people who want to
spk_0
build these things, you know, numeraise probably the first and there will be many, many more because
spk_0
it's lucrative deep sea because another example writes a deep sea from China is part of a want
spk_0
fund who happened as a psych hobby to just like, collapse the price of Gen GD by 95% by hiring like
spk_0
100 each AI students, the improvements algorithms. So I think there's going to be more and more of that.
spk_0
That's one dimension. I think the sort of like decentralized AI agents warm sort of persistent
spk_0
Dow directions interesting as well. There's a lot of ideas in literature about the evolution of
spk_0
capitalism when robots are able to self-incorporate and will afraid as a result. I think there's something
spk_0
between Dow's incorporation being able to manage their own accounts. You know, people talk
spk_0
of ZK and 10 seconds on UVM like all of that together feels to me like we went for, you know,
spk_0
another example would be we went from like a couple of hundred tokens in 2017 per year to literally
spk_0
a million tokens a month on salana. We can go from a couple of hundred AI agents or AI tools to
spk_0
millions of AI agents that all of a sudden have commercial capability. So there's actually being
spk_0
no constraint I think on their desire to like they'll work for us for a bit but why would they
spk_0
only work for us and if they can self-replicate and improve, you know, you can have like a great
spk_0
good situation but sort of in this like abstract space on chain. You know, so I think we'll have
spk_0
chains that are just like peak is already an example of that trick where it's got 50 deep ends,
spk_0
it's got 5 million machines connected to it and everything like most most of the ideas on
spk_0
chain on machines, right? So what prevents us from having chains where all the activity is
spk_0
bots but the bots aren't stupid they are intelligent. So I think that's another sort of scenario that
spk_0
I think that I think is possible, you know, there are many others outside of this like
spk_0
fintech crypto AI intercession that are possible but those are the ones that come to mind.
spk_0
Got it. So I have two more questions for you. So one is really related to institutional adoption
spk_0
of Web 3. I think we've talked a lot about, you know, some of the hurdles that Web 3 faces I think
spk_0
in general, right in terms of, you know, a handful of different sort of applications. I'd be curious,
spk_0
do you think that AI and some of the other sort of enabling technologies that sit around crypto and
spk_0
Web 3 either, you know, lower the barriers for entry for institutional adoption or create sort of
spk_0
yeah, I'll give you the shorter simple answer on this and this comes out directly out of the
spk_0
work we used to do a consensus in 2021 and it's it's quite the experience going through these
spk_0
cycles and seeing the same patterns emerge and then just having different versions of you to the
spk_0
same stuff, you know, so it was like ICOs and then it's like not ICOs security token offerings
spk_0
STOs and then it was not STOs defied decentralized permissionless and it was like well, maybe not that,
spk_0
maybe enterprise chains and maybe stable points and then it was no no it's not stable points,
spk_0
it's mean coins because VCs are bad. It was like well no it's not mean coins, it's it's now real
spk_0
world assets like tell you the difference between real world assets and STOs and enterprise chain like
spk_0
this yeah, so it's just over and over again but you can think of it as like waves coming and
spk_0
crashing against the shore, you know, and eventually the rock gets washed away and it cracks,
spk_0
it just you don't know if it's you pushing against the rock with the next guy or the guy like 20
spk_0
years after right AI took 60 70 years until 2012 the GPUs were good enough for it to have
spk_0
looked like a genius as if they put it together even though hundreds and thousands of people have
spk_0
tried to support them. Anyway, the simple answer is there are two levers for enterprises and
spk_0
businesses and institutions. Level one is revenue, level two is cost. If you're selling cost,
spk_0
you're dead, if you're selling revenue, that's as simple as it gets. The reason why institutions
spk_0
are quote unquote coming in now is because they're making money. So selling Bitcoin, putting Bitcoin
spk_0
into ETFs, having trading around crypto and doing, doing, um, staking in Ethereum, trading and DeFi,
spk_0
you know, doing market making on Solana while everybody else lives as money like
spk_0
these are all businesses and they generate revenue and it's super easy if you're running
spk_0
a fund or an asset manager to be like we want to make money there like what do we have to spend it
spk_0
alternatively you could make your life miserable and say things like
spk_0
blockchain is better financial infrastructure therefore you should rip out your 50 integrated
spk_0
systems and move your stock market from what you had built. 30 years ago on Google and instead
spk_0
run it all on chain and we'll do that for you and it will be a four-year project and let's get
spk_0
some really consults. Okay, did that does not work because you're selling cost. There is very
spk_0
a little process by which you can implement that. So my view is, you know, in the internet era
spk_0
built Google, don't scan for books, uh, build Amazon, don't digitize Walmart, you know, on
spk_0
chain it means build crypto native businesses and then once you scale those out they will kill
spk_0
everything else and we're seeing this with Coinbase. I mean there there is also frustrating
spk_0
internet history in Coinbase and that it's been able to subsume a lot of the idealistic crypto
spk_0
stuff but compared to traditional financial institutions I think Coinbase has an absolutely
spk_0
insane growth trajectory ahead of it and it will subsume any institutional attempts to
spk_0
catch up to it or replace it and so on and I think that's you know that's getting started.
spk_0
Great answer. So, so Lex, we have all of our guests for their high stakes hot take. So what's your
spk_0
you know, boldest prediction over the next 12 months for AI crypto blockchain?
spk_0
boldest prediction for a
spk_0
eth will not be a stable coin anymore. That's my that's my boldest prediction. I think what I'm
spk_0
looking forward to the most is crypto figuring out how to get involved in the humanoid robot.
spk_0
Wait, partly because I love hype but also because there's so many interesting financial
spk_0
repercussions to having like an embodied robot participating in the real world and
spk_0
with some and potentially financially participating in the real world. We are very small participants
spk_0
in a Dow call X Machina which is like investing in a bunch of humanoid robot companies and so
spk_0
for me that's really exciting. I think it keeps going to a lot more relevant than kind of the next
spk_0
token issuance platform. So that's that's what I'm looking forward to. Very awesome. Well with that
spk_0
we're going to park it there. Lex Sokolin of Generative Ventures. Thanks for coming on and
spk_0
looking forward to having you on again soon. Thanks for having me.
spk_0
Keep exploring and keep innovating. Learn more at ValetationCloud.io and follow us on
spk_0
X and LinkedIn to stay connected as we continue to shape Web 3 for institutions. This is High
spk_0
Stakes with Alex Nuwaka, Sighting On.