# Network Tokens, Stablecoins, and AI: Crypto's Strategic Evolution

**Podcast:** web3 with a16z crypto
**Published:** 2026-06-12

## Transcript

As AI becomes more powerful, it becomes better able to circumvent constraints.
So how do we truly constrain it?
Well, cryptography.
Cryptography for identity, cryptography for authentication, cryptography for money.
Technology accelerates the production of mediocre things as a side effect of producing the incredible things.
People are spending like $5,000 a day on tokens but only paying $200 a month because the companies are raising rapid capital.
Like we're in the attract phase, not the extract phase.
Okay, well, what happens during the extract phase?
Yeah.
That's exactly where crypto is going to come from.
And I think it's just a complete no-brainer.
Stable coins are being adopted everywhere and in all the corners of the economy.
What we are talking about is crypto's first killer app.
They are not a full end-to-end fix of the whole chain.
No.
They're not.
They solve one specific part of that chain and they do that one little piece so fast, so cheaply, and so securely that...
Hi, everyone.
Welcome to the A6NZ Crypto Show.
And I am so excited today, first of all, to be back on the show after a while.
Excited to have our co-host, Robert, on.
But most importantly, our special guest is Eddie Lazarin, who I have known for a number of years.
But recently, he was announced as our newest GP general partner in the crypto fund.
Actually, what does it mean to be a general partner?
I think technically, in our case, it means having the direct ability to write checks.
to speak.
Yes.
And there's fiduciary responsibilities with that.
Well, I do know one of the fun ways that the fund operates is that it's all by somebody saying yes.
There's not a veto rule where somebody says, no, we're not going to do this.
It's all, if somebody has conviction, then they go in.
That's right.
Well, is it more precise to say it's not consensus-based also?
Because when you say that, it sounds like, like, basically, you don't need, like, group...
Yeah.
No, our investment is what some people call a single trigger.
A specific GP, a single GP can unilaterally make an investment decision.
Can someone kill a decision?
What would a non-consensus bet be in crypto today?
In a way, what's funny is how do you identify what non-consensus is, right?
If it's clear enough to be something that you can point at, then that's probably because some people have formed consensus around that and made it visible.
made it something you can interpret.
So I think in retrospect, things that are non-consensus bets become evidently and observably non-consensus.
But at the time, they're probably defying some kind of pattern.
This is where Peter Thiel's famous thing came from, which Chris Dixon actually talked about a lot when he came to the A, 6, and Z over a decade ago.
And it's basically the idea that good ideas start out looking like bad ideas.
You kind of have to find an idea that's really a good idea, but it looks like a bad idea.
And that was his definition of finding non-consensus.
Yeah.
And, you know, looking like a bad idea has many dimensions to it.
One can be that it's not consensus.
So that makes it look bad.
Right.
Another could be that along some critical dimension of the consensus idea, it's completely at the opposite end.
Right.
Of some specific part of it.
Right.
It's more obvious in retrospect and it's not a thing that we have a strong sense of now.
You know.
What's non-consensus in crypto?
I think just because something's non-consensus doesn't mean that it's good either, right?
Obviously, like, there's a lot of things that are non-consensus in the sense that there is consensus that they're bad.
And, you know, I hate to admit, like, some of those categories are probably on the weaker side.
today.
It's probably pretty obvious.
Well, your point is there's a McKinsey quadrant, right?
You've got the bad stuff that everybody knows is bad.
You've got the good stuff everybody knows is good.
But you want to go for the stuff that people think is bad, but actually is very good.
And that's a distinct skill to be able to figure that out.
Is it also possible that given that we're on the cusp, we don't know, but you know, the indicators are that clarity could pass.
that the seemingly bad things can become good things?
Because for the first time, we might see people that only did, quote, shit coins, meme coins, whatever they're called.
I don't want to be derogatory because there may be some community value, but they can actually evolve into something with the business model now?
Yeah.
I mean, maybe evolve is one way.
Maybe we just need a new crop of projects, right, that design themselves from the ground up to take advantage of what clarity lays as a foundation.
If clarity passes, then I think, A lot of people are going to go to the drawing board and think about what this enables.
Yeah.
And design completely new projects to be refreshed.
Part of this market structure legislation for crypto is distinguishing between tokens that are related to networks and tokens that are related to companies.
What is the difference between those two things and why does it need to be written down in law?
I mean, the Clarity Act is over 300 pages, I think, right?
So forgive me for glossing it a little bit and oversimplifying.
What a network here to me means is that there is a wide array of people or institutions that are all interconnecting with each other in some way that the sum total of their interactions is more valuable in this thing than individually.
And a network token is something that is the natural ownership asset for a decentralized marketplace.
That's the idea.
So there's a decentralized marketplace.
And if you can own part of that, we call that a network token.
So a blockchain is an example of a decentralized network marketplace.
It's a marketplace for compute, for storage, or for verified compute and storage.
That's what a blockchain is.
Uniswap or what used to be called Maker is now called Sky.
These are projects that are also marketplaces, marketplace for liquidity, marketplace for...
for leverage or for borrowing.
And because so much of those things can be fully automated, they can capture value in an automated way.
Because they can capture value in an automated way, and they can distribute that value even in an automated way, there ought to be an instrument that reflects that, reflects the ownership in the captured value and the ownership in the upside of that thing.
That's what a network token is.
Now, you can see how that's...
similar to an equity asset for a company, except there's a key difference.
I keep using the word automated.
A company may have components automated, but it's still under the direct control of its management, whereas a network token might be a part of a decentralized network that doesn't actually have direct controllers, direct management.
So it's just trying to get at this exact distinction about things that can be fully automated and these things that cannot be fully automated.
So tying it back to clarity, the big picture is that previously a lot of legitimate players like founders, crypto founders who wanted to build real businesses didn't have the incentives to do that because they were afraid that trying to have a business model for their token would mean they were suddenly classified as a security.
And then all these other implications, you could have people come after you for the wrong reasons, et cetera.
And so like what does what's the next step if this, quote, design space is unlocked?
If it's unlocked, people can look and say, here are the milestones or these are the conditions I need to meet in order to be a network token.
And that means that they'll be able to issue new tokens to represent new types of networks.
And they won't have to worry about the SEC deciding to classify them as a security.
I think we'll see a lot of experimentation by people who find clever ways to fully automate things that could not previously be fully automated.
It's framed around automation.
That's like really autonomous behavior.
Behavior without a controller.
The way to really think about it is just that computers can do more and more.
What's the limit of that?
What's the limit of that?
Do you want to be able to own a thing that a bunch of computers can do on their own?
I think you do.
I mean, I feel like abstraction has run our lives today.
So I think it's not a bad thing.
That's why I work here.
This is probably a really fraught analogy, but I always think back to the Leviathan cover from Thomas Hobbes, where it has all these people sort of assembling into one giant king ruling over the land.
That's kind of my visual depiction of a network.
That's funny.
I mean, I think the key here, too, is if you think about crypto blockchains and beyond, just from setting a definition perspective, automation...
in your definition, is sort of like a way of saying this thing can run without a controller, i.e.
a manager of the network.
But there is another nuance here, which is ideally post some good legislation like Clarity or something like that passing, people can actually still be involved in their projects.
Because in the past...
People abandoned their projects, thinking they had to show that they're decentralizing.
Yeah.
I mean, one of the key dimensions to define that something's a network token and not a security is whether or not there is a single controller.
It's not hinging around whether people are contributing or not, right?
It's not around continuing efforts.
Yes.
So you can keep supporting a thing that's still mostly or fully automated.
Just because the thing is automatic.
doesn't mean that it can't be upgraded or can't be supported or can't be improved or can't have things layered on top of it.
And I think every type of developing network constantly needs maintenance from its users and from its participants.
I mean, you could theoretically have a roadmap.
Yeah, which you couldn't for a very long time, right?
And you also couldn't have value capture.
And I mean, something I've really appreciated about the last year in crypto has been a resurgence of interest in revenue.
Right.
Is in projects taking revenue.
I'm a big fan of revenue.
I've said for a while that that all the people who said, you know, revenue is like a minus or it's actually a it's actually baggage.
Right.
It's bad.
You don't want revenue.
They're confusing many things.
Like it's like a combination of cope because if you don't have revenue, well, you know, yeah, the revenue sucks.
I don't have revenue, but revenue sucks.
Who cares about revenue?
Right.
But it's also- It's like the famous line in the show Silicon Valley where they're like, you don't want to make money.
You don't want to show that you can make money.
Yeah, yeah.
It's all the dream.
Again, I get.
The tractors will make arguments like, well, if you have revenue, then people can value you.
Or if you're taking too much revenue, then you're taking capital out of the system that could be used to growth.
These are totally reasonable complaints.
But the problem is, of course, is if you just defer revenue forever, then that That may be because you have to defer revenue forever because you don't have any pricing power.
So that's why I've been a big fan of decoupling pricing from growth spending.
So in a network token, I've been a big fan of do your pricing up front, test your ability to price in the market, collect fees in the market, and validate your ability to price the thing.
But then on the other hand, there's this idea of the buy and burn in crypto.
The buy and burn, a well-popularized, well-understood idea.
Another one of our GPs was a little bit down on this.
Yeah, which is great.
We argue about these things.
But the idea of the buy and burn is that you take revenue in one side, you charge whatever you can get away with in terms of your price.
You take that revenue and use it to burn the token, burn a network token, and reducing its supply, which helps it appreciate.
If you have a constant buyer that's constantly buying and retiring, that may lead to price appreciation for the asset.
Anti-deflationary?
Yeah, yeah.
It's deflationary, you could say.
And then on the other side, the sort of mirror of the buy and burn is the mint, the mint and spend, so to speak.
And this is where you are minting tokens out of nowhere to pay for whatever you need, for growth or for network activities, right?
Often when people are critical of buy and burn, what they say is, why are you taking in revenue?
First of all, you're slowing your growth by charging.
And then second, you're taking capital that came in and you're just burning it to appreciate the token instead of spending it on growth.
That's a totally reasonable complaint.
But one of the magic of tokens that's not really viable for stocks is that you could keep minting on the other side in a way that is totally decoupled from the fees coming in.
That's fascinating.
So in other words, you can keep pulling in fees and test your ability to charge from the market.
while completely separately scaling your spending in order to facilitate the most growth.
And you may end up at a point later where you're burning much more than you're spending.
Great.
Okay.
So you've reached a point of maturity where the defensibility and network effects are strong enough that you can finally take a little bit of value from the market.
But maybe in the early stages, you are...
burning a little bit while you test your pricing, learning about what the market's willing to endure.
And then on the other side, you may be inflating actually, spending to continue a cycle of growth, right?
So you can decouple these things.
Because in stocks, you can't be issued continuously to pay, like you can't pay an onboarding bonus to Uber drivers or whatever with your stock.
Stocks lack this capability to decouple these things.
To pay for growth, a conventional company has to do it in a very indirect way.
They have to sell the equity to an investor in a private round to get cash to spend on growth.
But this allows you to circumvent that completely because as long as people can actually invest in the token, then it has a market value, which is the thing that you're observing over time as you mint out of one side and burn out of the other side.
Just to put a finer point on this, Eddie, does the token matter here?
Like, are you collecting fees in stable coins and another sort of in a network token?
And like, what kind of token are you collecting?
Yeah, it doesn't matter at all.
Yeah, it doesn't matter.
And then you're minting in your own native currency.
It could be whatever you want, right?
Yeah, revenue is revenue.
Revenue is revenue.
You know, does it matter that an American company gets paid in, you know, Brazilian reals or whatever on the other side?
Like, it doesn't matter, right?
I mean, there may be an extra swap step, you know, some sort of...
fee management, capital management, foreign exchange style thing up in front, but in the end, like revenue is revenue.
So the receipt of revenue is less important in what form you're receiving it in, but the minting is really, that's going to be your own native network token, right?
Yeah.
So going back to the Clarity Act is experimentation around these types of things.
How do we capture value?
How do we price the thing?
How do we think about, from a market's perspective, valuing a token?
How do we think about valuation?
All these things have all been bouncing around in crypto for some years now, years.
But because we lack the Clarity Act, we haven't really been in a position for the market to do its experimentation and to learn about what actually works and what doesn't work.
I'd love it if a few years from now, we found out, you know, which parts of what I just said don't make any sense and which parts are great, if there are any, right?
Like, I would love to know that.
But we have...
Not really experimented in that for reasons that make sense.
Entrepreneurs don't want to take extreme risk with their project based on a murky interpretation of some old laws before the new law passes.
So I can't wait to see what the market tells us because I think of the market as an information-producing machine.
It is constantly doing this.
And it's so interesting sometimes to step back and look at what the things are that the market is willing to spend resources to learn about and which things it finds too scary.
Yeah.
I have two questions here.
So what's really interesting about what you just outlined, I've actually never heard it altogether that way, is you're talking about what's super unique about tokens and network tokens, this property they have that's kind of not ever been seen in history before.
It's a new mechanism, a new mechanic.
Similarly, though, there is something very traditional about it, which is very refreshing to hear, which is you're actually asking people, hey, charge, have a business model.
For sure.
Figure out how to make money, charge fees, like don't just grow with like fake incentives because tokens have this quality.
Like you actually have to find product market fit this way, like instead of just.
Fussing around.
Absolutely.
Yeah.
I'm not fully bucking the sort of Silicon Valley stereotypical view that, you know, growth is all that matters.
I'm actually agreeing with that.
But I'm saying that you can experiment with where to do pricing and where to create incentives.
Basically, if you have your own token and it's part of a decentralized thing and you can rally your community around.
this type of thinking, you can engage in essentially like monetary policy style behaviors.
Exactly.
Yeah.
Which is not really possible for companies today.
Companies don't think of their equity as something that they can experiment with too creatively.
Of course, they do through fundraising and things like that.
But beyond that, it's not this sort of fluid experimental thing.
And I just, like something I've learned about markets is that.
They love liquidity and capital loves like sloshing around to where it's safest and where it can grow.
Crypto offers that.
It offers a very high freedom to the capital to move around and reallocate itself.
I refuse to believe that the market won't value this eventually.
Maybe not right now, but the market will value this.
Well, then this goes to my question.
What would you say to the people, whether you agree with this or don't, that would argue that crypto's only use case is speculative?
Because I think a lot of people think, well, liquidity is enough for liquidity's sake.
And the ability to trade on it.
build derivatives on it, do whatever you need perpetually or otherwise is a unique thing that crypto offers.
And that's where it ends.
Yeah.
I mean, I admit to some degree that that has worked for crypto.
Yeah.
Crypto has been very much about speculation.
But take a step back.
Like what is what is speculation?
Speculation is where you're kind of allocating some money around and you're kind of seeing what happens.
And I guess the difference between mere speculation.
right, in the sort of casino or gambling sense versus speculation in the capital markets, you know, things that are actually growing and creating value sense.
It has a lot to do with the types of things that the capital can flow to.
Because if the capital can only slosh around inside its little kiddie pool, well, then it's going to be speculation because there's nothing concrete that the capital can be anchored to.
Because think of the capital as like liquid.
human effort or something.
That's a great way to put it.
Think of it like that a little bit.
Then if it's just sloshing around a little pool, then there's no way for it to be absorbed into a project to create more productive efficiency, which is how capital grows.
It grows by...
concentrating into like a little machine that like unfolds and then makes something more efficient or more useful or, you know, better priced or faster or more secure or whatever, right?
It's like this improvement.
It needs to be able to connect to the rest of the economy to do that.
So as long as crypto has been walled off, sort of firewalled from the rest of the economy, it's actually kind of obvious why it has failed to take root in things that are at the full level of productivity improvement that we imagine they could be.
Like we used to say years ago, it's still true, is that, you know, like meme coins are illegal.
It is legal to experiment.
Why is it legal?
It's legal because it should be very obvious that they're inconsequential and that their price action has little bearing to do with reality and the efforts of an entrepreneur or a company.
You know, like it's kind of obvious that it's unmoored from reality.
The minute you start talking about getting moored to reality, you reenter a massive, you know, regulatory thorn bush that you need to respect profoundly.
The Clarity Act basically creates the conditions that this capital can connect with reality.
And the Genius Act already did part of that.
It's gone from like a thing we were noticing to like a thing we are fully swamped under is all the opportunity that the Genius Act unlocked.
Yes, with stable coins.
With stable coins.
And that's because...
It created the very specific and concrete conditions and excitement for how those conditions will develop between financial institutions and real capital-related and money-related use cases.
Because there's that link now, everybody's flooding around it because the real world can connect with the crypto world.
So if we get the same thing with Clarity, then we get maybe, hopefully, the same outcome where people can say, wait a sec, now market structure, now capital formation around crypto-native assets.
is possible.
That's what I'm hoping the next year will be like.
I think sometimes we put a footnote on stablecoins, like it's just a passing thing.
And I really feel we're underselling the opportunity, quite frankly, like and what the actual phenomenon is like.
What we are talking about is crypto's first killer app, which we talked about in our state of crypto report.
Robert, you know this better than anybody.
But this is actually exactly where Degree is taking over and can very quickly tip into kind is.
Stable coins are being adopted everywhere and in all the corners of the economy.
It's not like crypto insiders.
It's TradFi.
traditional companies, international players.
It's like massive, the kind of growth and adoption and primitives we're seeing and what people are doing with it.
And so I just think it's important to highlight that that might be another one of these examples where the sheer degree of it can manifest something bigger than what we think.
Stablecoins, I think, are, even though they have become a consensus topic in crypto, they've even become arguably a consensus topic in fintech.
I hear from a lot of Yes.
Conventional, non-crypto fintechs or whatever are talking a lot about stable coins.
I mean, the way I like to give this story, maybe I'll just give like why it's so, why they're so impactful is like, basically when you want to move money in one part of the economy to another, whether it's in a developing country or in the United States, it has to go through a series of intermediaries.
It has to go through a chain of intermediaries.
And each of those intermediaries is not optimized for speed and cost.
They are optimized for compliance, for adherence to rules as they exist.
And they are also just kind of barely doing as much as is needed to continue to function in that role.
That is my impression.
Whereas what stablecoins achieve is they are not a full end-to-end fix of the whole.
No.
They're not.
Instead, they solve one specific part of that chain, a very specific part, which is literally the moving of a balance on one side from one address to a balance on another side.
And they do that one little piece so fast, so cheaply, and so securely that it is like the entire rest of the system.
is right now trying to reconfigure itself around this one little Lego brick.
It's like, imagine a circuit, you know, there's like this tangled web of stuff and then there's like...
This one little piece can do like a gadrillion per second at like no cost, basically.
And connecting to what you said earlier about essentially arguing that capital flows is innovation.
What you're basically describing, and you described those earlier capital flows as water-like.
You're describing a system where water flows anywhere and everywhere in a way it can permeate through various boundaries.
It can seep through pores.
It can break down things through attrition of the rocks, whatever.
It's literally doing that in the financial system and beyond.
Absolutely.
Capital wants to be free.
Yes.
And it will do whatever it takes.
Like information used to want to be free.
It will do whatever it takes to get there.
And so right now it's reconfiguring itself along these lines.
And that can't just be done in one shot.
There's many, many, many, many little pieces that need to reconfigure on this, specifically to exploit this speed and this savings.
Yes.
And what will happen as they...
shift over and as new companies form and can do the end-to-end, they will have many second-order consequence improvements because of this.
So, you know, add to a bank, very simply, that it can send and receive stablecoins.
So you have a balance in the bank and you can withdraw it in stablecoins and you can deposit with stablecoins.
That's all you do.
You just take a regular bank and you add that.
A bunch of things become possible because now that they can send and receive stable coins, well, now instead of your balance getting withdrawn from the bank, they could take your balance and make it in its own address or its own wallet.
Okay, well, then that means someone else can see your collateral without having to access the bank.
Now that means that someone can lend you something because you could make it possible for...
for people to see things going in and out of your account, maybe a simpler example is you can take two different banks and you can connect those two different banks because they both speak stablecoin.
Yes.
So now these two different banks can talk in a way that would have taken a bespoke integration.
Now they're all connectable with each other.
Yeah.
You can take all kinds of products and you can weave them together and connect them all together.
Exactly.
They can all sort of speak the same language instead of today, which is they have to do one by one by one integrations.
And that is...
Again, not a change in kind.
It's a change in degree.
Yes, but a very profound one.
But it ends up in a, it's all whole new kind qualitative difference.
Exactly.
It becomes that.
In short, I think crypto is entering a different phase.
This is basically exactly related to what we were talking about earlier, which is that we've alleviated many of the technical constraints.
Exactly.
And what are examples of those technical constraints?
One is...
blockchain throughput as a result of improving consensus algorithms.
Some of the throughput was constrained by consensus algorithms needing to improve.
That whole space has advanced profoundly, right?
Profoundly.
We have so much more.
I don't know the latest benchmarks, but I suspect hundreds of megabytes per second in consensus, or that's just ordering capacity, which can represent gigabytes per second of data, depending how you quantify it.
But blockchain consensus has improved radically.
snark design and the applications of zero knowledge have improved radically, or again, orders of magnitude over the last years that we've been working on this.
Those things will continue to advance, and it's important that they do.
But like four or five years ago, when we thought that the constraints were purely technical on research, that became a big focus for us.
And it remains a big focus.
I think it will be a constraint over time.
Just because of where the space is today, I think we think of the constraints as product design, as regulatory.
And I think of as adapting to the market where it is.
Like that's maybe part of product design, but it's like all the crypto enthusiasts who just wanted to own crypto assets, they're already technical, do stuff.
They have everything they need.
And they have had everything they've needed for a couple of years now.
If you do the extra work and you write a little bit of software.
You can run your whole life basically out of crypto.
I do because that's kind of my job is to suffer through these things.
So I definitely do basically everything.
But that's not what we're in crypto for.
We're in crypto to reach the mainstream, to completely modernize how the entire world, normal people, the most sophisticated institutions, the cypherpunks revolutionize how everybody handles finance.
And handles money and handles capital, which isn't going away.
And so we are short of that goal.
And so I think the entire industry is trying to adapt and figure out what the constraints are.
Because there's just no way that it will remain as it has.
I mean, it already is adapting.
It's exciting times.
Yeah.
I mean.
So that's like deep in the weeds on how network token business models work and legislation and all that.
I would really love to use this opportunity to understand you better, Eddie, like why this excites you in particular, why any of this excites you, you know, where you come from as a person.
When I think about A16Z crypto, the culture of it is very much set by its founder.
Chris Dixon, who everybody regards as a philosopher, given his degree in philosophy and his continued interests in that.
And one thing that excites me about your elevation to the post of GP is that I see you very much in that same vein, that you are steeped in philosophy, always have been, have a degree in it.
And that's why at every offsite...
I make my beeline to you to get more reading recommendations and understand what is the latest thing on your mind.
And I'd love to share that sort of intellectual enthusiasm that you have with the rest of our audience here so they can understand, you know, the kind of person that you are and where you come from.
By the way, I would call him more of a philosophy engineer, too.
That's like his identity.
Yeah, I accept this.
I accept this.
I accept this label, Robert.
I'll accept it with a caveat, though, is that I do find it can be a little bit cringe to me, a little bit, because it's often unclear what people mean when they say this, right?
Because sometimes they just mean a philosophy-ish person as like a...
A non-doer?
Yeah, yeah.
A ruminator?
Yeah, maybe something like that.
A non-doer.
A yapper?
Like a Hamlet type?
Just like thinking about thinking about thinking?
Yes, yes.
Which is excessive...
That is cringe.
The excessive ponderousness is like...
Thank you for being self-aware.
This is what I love about you, though.
You're actually self-aware about that cringe factor.
You're defeating your own point.
Clearly, you've thought through this too much.
Yeah, exactly.
Well, but in a good way.
He's self-aware.
What I think people mean by that with the philosopher angle is just like really trying to understand what it means in the most basic terms, like trying to root it as deeply as we can in all of reality a little bit.
Okay.
Like this conversation, we just had a little bit about like the network tokens and like value capture and all that.
this type of thinking becomes a lot clearer when you try to think about what the market is and what's happening in the market.
So let me just say like a couple of things about how I think about the market as a means of introducing myself, Robert, because you actually asked about me, but I hate talking about myself because I see myself as a conduit for the market, right?
I want to be a conduit for the market because it's so- Oh muse of capitalism, speak for me.
Yeah, I want to understand it because it's so- It's teaching us so much.
When we talked about network tokens, we talked about value capture.
Value capture is a thing that the market is obsessed with.
Now, here's a very unconventional opinion.
Okay.
Okay.
Okay.
Very few people agree with this, but this is like a fact to me.
The reason why capitalists and venture capitalists and so many business people and entrepreneurs are so obsessed with value capture is not because it's like serving their greed.
At every turn.
Not because it's like, you know, how I get rich, whatever.
It's actually because value capture is incredibly difficult.
And the market is leaking its value at an unbelievable pace at all times.
Almost all the value that's created in capitalism is leaking at full speed.
There's this idea, something called Schumpeterian profits.
Okay.
Schumpeterian Profits.
There's a great 2004 paper by William Nordhaus that's very readable, really nice.
There's a whole literature around this.
That whenever a business improves its productivity, makes products better, makes products cheaper, you know, whatever it does, better.
What percentage of the improvement, so to speak, its fundamental improvement in productivity, is captured by it in profits?
Well, according to Nordhaus, what percentage would you guess of the...
This surplus, this sort of improvement in value, is captured by the company versus captured by everyone else or everyone else.
I would guess very small, only because you talked about this.
I led us to the answer.
The leaky capitalism.
According to Nordhaus, and I'm sure this has changed a little bit over time and varies from sector to sector, but according to Nordhaus, 97.8% of the value is externalized.
Wow, that's significantly wow.
Very little of the value is captured by the company.
Also a big narrative violation, by the way.
A huge narrative violation, right?
Well, certainly in areas where technology has the ability to like crash prices, right?
I mean, that would make a lot of sense.
Of course, and replace things that are more expensive and make things last longer and forego the need to even purchase this thing.
Capitalism does this all the time.
So take the fact that the value is leaking all the time, which means it's very hard to...
keep the value in.
The reason why the value is leaking is one, competition.
The obvious one is that if my company is making things better and faster and cheaper, I want to take all that from you, your company, so I lower the price and no one's going to buy from you.
And now you lower the price and we kind of go back and forth and I raise money to invest in an even better system.
Only the consumers benefit.
The consumers benefit.
So that's a very obvious place where the value is leaking.
And by the way, every price that's posted is an advertisement for other competitors to jump in.
If you're charging a high price and a competitor sees that, that's just like a blinking, blaring red signal sign for them to jump in.
The entire market is sharing this information at all times to encourage everyone to compete with everyone else to take all the value out.
It is the beauty of market capitalism, yeah.
It's constantly leaking.
So why is there such an obsession on value capture?
Because first, it's rare and isn't the norm.
That's usually not what happens.
In the end, all the benefits diffuse, whether it's new knowledge, whether it's the actual surplus from the investment.
I love this, yeah.
It's constantly leaking.
Think of capitalism as this ship, and it's just sinking all the time, but we're building it faster than it's sinking.
It's actually in a runaway feedback loop where it's growing and leaking at an accelerating pace.
It's leaking faster and faster, and it's growing faster and faster.
It's a perma-boom, perma-crisis.
at all times.
And the second reason why people are so preoccupied with value capture is because value capture is very visible, very legible.
Because you can see it in the company, right?
You see profits, you see the costs, it's on the accounting sheet, like it's the easiest thing to track ever.
Whereas the surplus you created for your consumers, it's almost impossible to even measure.
The extra work you'd have to do to measure that I gave you a new good.
That's way better.
Three other things that you used to have to buy, you don't even have to buy anymore.
And this thing lasts longer and whatever, whatever.
Like that's very hard to quantify that.
Really hard to quantify.
I mean, how do you quantify that?
Like you bought a new thing and you love it.
Okay.
You love it.
It costs 15 bucks, but you love it.
Who is taking the accounting?
Like literally who is measuring that you got, like you would have paid 200, but you only had to pay 20.
No one's accounting for that.
In my view, the vast majority of the benefit is constantly exploding out of it.
And the struggle of the entrepreneur is to struggle to keep enough of the benefit inside that they can keep feeding the machine.
Yes.
Capture enough value to continue.
Because if it bleeds out too much, it just dies, right?
So it needs to be able to keep its blood inside of it, right?
And that is an incredible struggle because it's not just that people think of it wrongly like you're fighting with the consumer.
No, it's you versus like every other company.
Exactly.
And every other company is, you know, fighting and trying to grapple with you and stab you and take your blood out.
And you're trying to keep it in your body.
That's what it is to be a company in capitalism.
So going back to me, Robert.
Philosopher.
It's just like the market is not a thing that seeks equilibrium.
It's a thing that is constantly in an accelerating feedback loop where there are like.
accelerating and decelerating factors impinging on its constant sort of barely controlled explosion.
Looking at the world like that, that's not very technical and specific from an explanatory angle.
But if you bury that in the back of your mind and you return to that, then when you do get into the technical details of how things work, you can sort of reorient yourself against the sort of frame.
and allow other things to kind of click into place.
Like many people have said things like this, like, you know, inverting the frame all the time, like, you know, the Jacoby thing, like invert, always invert, always invert.
Yeah.
These sets of tools to use the Chris Dixon phrase, like mental models, right?
He's always about his mental models.
Having a great toolkit of mental models is what a generalist does.
It's also what a philosopher does.
Yes.
Right?
And that's how you can encounter totally new domains, many, many new domains, and very quickly get to just a couple clicks deeper than an average person vantage point.
Because that is, I think, the ideal role and the ideal specification, I think, of a venture capitalist is not to go deep and be the deepest expert on every single thing.
You can't possibly.
That's the role of the founder.
But to be a competent intellectual partner and sparring partner for founders so they can discover you and you can discover them and you can make them a little wider and they can make you a little deeper.
That's the perfect distillation.
So you've listed one person here, Nordhaus, as his ideas being influential to you.
But who do you count among your greatest influences beyond him?
Yeah, think through your intellectual upbringing.
Who would you cite among the pantheon?
And it could be people you also used to disagree with and further form your views.
Absolutely.
That's a really good point.
That's especially good.
Who I disagree with.
People you hate.
You just can't dislodge them from your brain.
That's a great one.
I mean, so top of mind for me is like some of Land's ideas, Nick Land's ideas.
Yeah.
Big fan of.
Yeah.
The thinking about feedback loops and cybernetics and artificial intelligence and capitalism being the same thing.
Very big fan.
I particularly like Vincent Le's new book about Nick Land.
I quite like Sub Vincent.
I just saw him on a podcast recently.
Very, very, very good.
Very interesting.
Love his stuff.
In a similar vein is Marx, right?
I mean, that's kind of answering your question.
Like some people remark that...
Karl Marx was like the earliest and best cybernetician, specifically because he was obsessed with these feedback loops as part of like his, you know.
Capital circuit.
Yeah.
Very famously.
Materialist dialectics is kind of the process, like this kind of thing.
You ask like kind of love, hate, like.
Yes.
Marx is like ridiculous, like labor theory of value, humiliatingly dumb, like so many things that he missed.
But still some core ideas, like almost for comedic value, like Zizek's sort of traditionalist or, you know.
I forget what he has sometimes called himself, like a conservative Marxist or something hilarious, ridiculous.
There's some hilarious clip I love of Zizek saying how much he loves alienation.
Because he wants to be left alone at his house.
Yeah.
Just like not deal with human beings, whatever, which is like obviously alienation is this like traumatic thing in the Marxist frame.
But like I think in a contemporary frame, it's actually almost a welcome.
Like we kind of welcome our ability to sort of not have to engage with other people as a result of the markets intermediating us from different people all the time.
I distinctly remember the first time I read through like summaries and, you know, compilations of like Adam Smith and David Ricardo.
And like talking about comparative advantage, blowing my mind, right?
And that's something I thought of before, right?
Like just like so counterintuitive, so interesting to lodge in the back of your mind, always thinking about comparative advantage, specifically with regard to how, what it implies about the optimal allocation of resources in complex economies.
For example, like this is where I go into, I've told you this phrase before, like I've told Robert this phrase before is like, getting a little schizoidal about things.
Remember this, Robert?
I'm going to say it.
Say it.
Hit me.
This is obviously highly reductionistic, but you think about schizotypal thinking versus autistic thinking.
This is totally not scientific, what I'm saying.
They're metaphorical labels.
But schizotypal thinking as everything is connected and autistic type thinking as everything is distinct and separated.
These two extremes.
What's the stereotypical Aspergian or autistic person, no, these two things are totally different.
They're totally different.
You know, whatever, right?
Like, you know, like that type of like, you know, being able to understand like the pitches of different notes and exact machines and this version of this thing, whatever, everything is separate.
And then schizotypal thinking, everything's connected.
Like this is this and this is, they're the same thing.
And like, you know, this kind of like, it's all, it's all part of one plan, whatever, that kind of thinking to the same vein.
The debate today about AI is very much about the replacement of the human.
That's a thing of great anxiety.
But take comparative advantage thinking.
I was just saying comparative advantage.
If you've really lodged comparative advantage into the back of your mind, it should be obvious that even if artificial intelligence exceeds human capabilities in specific domains, there will remain work for human beings to do that even if human beings are less efficient than the most efficient AI, it makes a lot of sense to allocate that to them as opposed to machines because of some differences in their capability.
Or the comparative advantage of human beings.
Yeah.
So that can only occur to you if you lodge these types of ideas into the back of your mind and you're always thinking about them.
Like they're just kind of erupting from things schizotypally.
So I think of my reasoning style as very...
Schizotypal.
I love that.
I was relating.
Do you relate to that one too, Robert?
Or are you on the other camp?
Because I was relating to the schizotypal too.
I was thinking of Veronica Roth's divergence.
I'm a connectionist through and through.
I'm always, yeah, I'm making weird analogies.
I think that's more typical.
I think generalists tend to be a little bit more schizotypal.
Yes, you kind of have to be.
It's also how you find meaning because if you're not going to be a specialist very deep, this would be like the autistic end of the two models you outlined.
You almost have to find meaning by connecting things and seeing that, you know, at Park, at Xerox Park, we used to call this not just T-shaped people who are both broad and deep, but it's like a comb.
You have one whole comb and each of these spokes are all the areas of interest you go in.
So you're in the comb.
And that was a unique kind of profile, actually.
It's a funny thing, but it's actually they did like kind of an informal study of who are the people?
Because, you know, it had a reputation of being repeat innovation.
Like they've invented a lot of things.
that was the type of person that was there.
It was like a comb-shaped person.
Yeah, totally.
Like very schizotypal.
I'll also say like Hayek and his whole lineage of ideas.
Yes, I'm very Hayekian.
I hope like my articulation of the market, like it should sound that way, like to anybody.
Actually, we do want to hear a little bit about your evolution to who you became, who you are from the career side.
Yeah, I loved music.
I always thought I'd be a musician.
I used to play a lot of guitar and a lot of drums.
Oh.
You know, I was originally, when I went to college, I was originally accepted as a music major.
What?
I never knew this.
Did you know this?
I did not know this.
I also came in having taken...
like, you know, AP economics and loved it, loved it.
So I took some economics courses and it was very soon after that I switched over to philosophy and economics.
Oh, so those are your majors.
Actually, that actually not makes perfect sense.
Like you're one of the few people whose majors match your personality.
It's like you came out of the system, molded the way you meant or you molded it.
Yeah, I think I was like one or two classes short of like actually an economics major, but like I definitely had a minor or whatever.
I love music, but I have to admit, like I was actually last week like lamenting that I don't, I really don't have the time or don't think that much about music anymore.
Yeah, I think the way music is produced and consumed now has just changed a lot.
And I guess I kind of deliberately kind of gave it up because at least the way I used to like to listen, there was nothing as painful as like being out of the loop, like being half in the loop, if that may be.
I know, I know what you mean.
What I mean by that is like, I love to.
have a comprehensive view of everything.
Like I wanted to know all the new releases, all the new, like if I wanted, I loved the feeling of just the wave of new stuff.
And just whenever there was like a new thing, I was anxious, like, I can't wait to hear it.
I just want to hear it.
I always put it on and just be like absorbing the new thing and thinking about it and sort of logging it in my mind.
searching my mind for my reaction to it.
How is this making me feel?
What does this mean?
What does this mean for the history of music?
You know what I mean?
All those types of things.
And doing a half-assed job of that is brutal because you're constantly reminded of that your view is incomplete, basically.
It's incomplete or you missed a reference to something.
I don't know.
It's too painful.
So I very much almost intentionally kind of like gave it up.
By the way, this would be ideal for agents because I've been actually trying to make some agents that can do this with literature for me in a different vein.
Like you could set...
up a bunch of agents that literally do this, outsource this thinking for you?
Possibly, but I guess I think the joy of music is in consuming it synchronously, so to speak, to use computer language.
Whenever I watch YouTube videos or listen to audiobooks, which I do all the time, I always listen at one and a half to 2x speed.
Music is the only thing you can't do that to.
You got to listen to music one second per second.
Yeah, one to one in the moment.
So how did you go from being a music...
person in college, philosophy and economics.
And then I actually don't know where you went next, because all I know is that before you came to A6NC Crypto, you were doing data science at Netflix.
So what kind of took you on this journey?
Well, I guess just in brief, the more economics I was doing, the more I was veering into what was very trendy at the time, which was behavioral economics.
Yeah.
Very trendy.
Kahneman and Tversky-esque stuff.
Yeah, very, very trendy.
And so I joined a behavioral economics lab, which was excellent.
Such an enlightening experience.
But needless to say, one of the bottlenecks for a behavioral economics lab is the analyzing of experimental data.
We're running experiments all the time and you need to be doing statistical analysis.
So the PI at the lab I worked at needed more.
So he told me, go take a computer science class so you can actually do better on the analysis.
And so I took one and it was like a mind-blowing experience, which I want to say that kind of a brief funny remark, I guess, is that even when I was in college.
The financial crisis happening sort of toward the tail end and then the dot-com bubble sort of in the beginning end and also my upbringing in Miami meant that the image people had imbued in me of computer science was that it was kind of a glorified IT job.
When I say IT job, I mean like that it was like making sure people's computers were working, not to denigrate IT because it's incredibly valuable, but just that's how people talked about it and I'd worked at like a retail computer shop.
summers and stuff, like basically the equivalent of a Best Buy sort of.
So I kind of avoided computer science because I incorrectly thought that this is like what it was, right?
Which is ridiculous in retrospect.
And then I took this computer science course and it just kind of melted my brain completely.
I was just, I could not believe how cool it was.
Oh my God.
I could have so many stories of like seeing the professor teaching sorting algorithms by like instructing students to move.
according to like the rules of different sorting algorithms.
Like students in a line and they had to like sort out, you know, like here's a merge sword, here's a quick sword, like at the end, right?
Here's insertion sword, here's a bubble sword, right?
The fact that you could use abstract rules and that the work of computer science was in abstractions, it wasn't in the weeds of computers.
Of course it can be, but it has to do with thinking systematically.
It was incredible.
I remember that summer.
the Department of Defense did a demo.
This was in 2009.
A demo regarding computer vision that was taking security camera footage at an airport and then trying to classify people in the airport based on the suspiciousness of what they were doing.
Yeah.
So we would draw like a box around them and like, this person's not suspicious.
And then like someone stops and starts like rifling around their bags and it was like.
suspicion increasing, right?
It was like literally like the color of the bounding box would like become blue to more red when they're more suspicious.
I remember thinking it was a totally mind-blowing idea that a computer could develop a belief about what suspicious meant.
Yes.
Now we kind of take that for granted right now a little bit that they can have quasi-subjective views.
But at that time I was like, I cannot understand.
how a computer can have an opinion about what suspicious means.
I was like unable to grasp, like wrestle with this idea.
And it meant I was utterly compelled to switch.
basically to computer science, spend the entire last year and a half of my undergrad on only computer science, and then stay and do a master's degree in computer science.
And there's like, I could only think about this.
My goal was to understand enough that I could at least have a glimmer of a sense of what it would mean for a computer to have an opinion that like this image is a...
suspicious person or this is an anomalous event or whatever.
And yeah, so that's how I got into computer science.
Did you ever deliver that helpful statistical analysis to the professor?
Oh, yeah, yeah, yeah.
No, I stayed working at that lab.
Yeah, I stayed and I did a bunch of stuff with them the whole time I was doing my master's.
It was awesome.
There was a really interesting point because it was philosophy and economics.
It was so crazy.
I was taking an advanced class in statistics.
And at the same time, I was taking a philosophy class.
I believe it was dual listed with the math department called Bayesian Statistics and Causal Inference.
Of course, because that makes perfect.
It's a mathematical problem.
And it was as much it was taught by a philosophy professor who happened to have a great math background.
Yeah.
And then I was intersecting this, of course, with like computers having beliefs for a computer, of course, especially at that time, you couch beliefs in statistical claims, basically like.
Yes.
It's trying to reason about the likelihood of things based on its prior probabilities of other things.
And it's very much a statistical claim for a machine to have a belief, right?
And so I remember taking both these classes at the same time and just thinking like everything's coming together.
All these things are interconnected.
The schizoidal part was really connecting everything.
Yeah, ultra schizoidal.
I was just like, this is insane.
Yeah, so that's how I got to computer science.
They do connect though.
And I think it's actually, I think a big reason why people are in crypto.
A lot of people ask me, like, how can you go to covering just crypto after you covered, like, everything else?
Bio, fintech, you know, natural language, like, physics, whatever.
And it's like, because crypto is everything.
It connects all these things.
Economics, philosophy, psychology, like, you know, all these mechanics.
It's super interesting.
Yeah, it does.
Yeah.
Yeah.
And, you know, around that time was when the Bitcoin paper came out.
I was actually I was in the middle of my master's degree when people really started talking about it.
So it was actually an object of interest, like to professors.
And like we would talk about it.
I got very much into it around that time period.
I mean, there wasn't a lot to get into.
It was like Bitcoin.
You know, you buy it or sell it, you know, I mean, whatever.
Right.
But but still what we now know of as Nakamoto consensus and this idea of networks that can be.
Byzantine fault tolerant, right?
Tolerant to internal disagreements about the state of the network from different participants in the network.
It was a very, very interesting family of ideas.
So I got into it around then.
When I moved to the Bay, I wanted to be a software engineer.
I wanted to write statistical Python.
You know, that was like what I really wanted to do.
And all those jobs were in the Bay.
So I moved to the Bay, worked at a startup, learned a ton, just doing a lot of social network analysis at the time.
on what is now X, right?
And stuff like that.
But all my friends at the time were all crypto people or Bitcoin people.
And of course, crypto has evolved so much.
It's kind of hard to believe it's been so long.
It's been like 15 years since then.
You've talked about how you had your entree into crypto and into computer science, but how did you go from there?
And then you made some stops in between at Facebook and Netflix before you arrived here.
So just Bring us up to the present.
In brief, like at the time in the early 10s, the growth of social networks in particular, but the growth of the web and all the businesses built on the web meant that now many, many, many companies had a lot more data to store.
So there was this huge anxiety at the time about big data.
Oh, yeah.
I remember big data.
Oh, God, I haven't even heard that phrase in over 10 years.
I know, I know.
All these cool technologies were built around this time to deal with big data.
Hadoop and Spark and Presto and all this stuff.
So I was very much in that world.
I would say in the combination of like what became data science, data engineering, distributed systems engineering, trying to figure out how to help these companies.
absorb as much data as they could and do interesting things with it and do very interesting analyses.
It's funny, I say with only a little bit of lamentation in my voice is that that has basically been solved.
Like that whole data engineer war.
Like, you know, you can do incredible stuff today with just a subscription that would have taken like a whole team.
10 years ago to figure out.
So it's kind of done.
And so, of course, all the people I know from that era, myself included, don't think about that at all anymore.
Around that time, though, 2018, I was so interested in crypto and I really felt like crypto was the place to go, in particular because I was hearing about smart contracts.
And that was like a whole other level of thing.
And I looked around, I was talking with a lot of people.
And I met the A16Z team and I spent a bunch of time with Ali and with some of the team members at that time and then Chris and, you know, decided to join.
And the thought was it would be helpful to have somebody with kind of a data and statistical, you know, data science background who was also from a software engineering background to support investing.
Now it's a little bit of a bigger team.
It has been quite a ride the last seven years.
So one thing is when I think of those early days when you joined, it feels like a lot of the conversations on the crypto team, yes, there were interesting ideas around the future of media and different things that could evolve later.
But it feels like a lot of the early conversations were all constraint-based.
You know, like here's what you can and can't do.
And then suddenly...
It seems like I feel like the first year I joined the crypto team full time, it was all data availability and only this and only that and throughput and transaction limits and da da da.
And I feel like we don't even talk about those things anymore.
Yeah, I think I speak for many people in crypto when I say we believed for a long time that the space was limited by technical constraints.
Yes.
And it was without solving those technical constraints.
Many things that we've wanted to happen.
or have happened, couldn't have happened.
But I think something, and I think this is part of why in certain areas of crypto, people are pretty bummed, is that the technical constraints lifting has not unleashed a lot of the innovation that we wanted to unleash.
Some of it, I genuinely believe, is constrained by regulatory issues, which like the Clarity Act, which I believe.
Others, I think, have been constrained by you know, honestly, a little bit of a lack of imagination where people have been trying to concentrate on what seems to be working.
There is so much interesting stuff happening now, in particular as a result of AI, that the threshold, the bar that something needs to exceed to be worth taking a risk on is high.
I mean, think of it like this is like, it's really obvious there's going to be some great stuff in AI.
Oh, yeah.
Incredibly obvious.
And there already is and there will be more.
That just raises the bar for whatever idea in crypto.
What do you mean by raises the bar?
Well, just you're a founder.
You can choose.
Should I go to AI and do some amazing thing or should I choose crypto and do this other thing?
That raises the standard for what you might even be willing to endeavor, to risk doing in crypto.
And so having evidence that some things have worked or some patterns you can match against makes it more attractive.
So I think in crypto, the releasing of the technical constraints has not quite led to the explosion that we thought could happen.
But often it's a matter of timing and certain things just take a while to unfold.
For example, in AI, like the AI takeover that we've seen has been since when?
The very end.
Chat GPT.
Yeah, like maybe a year after that.
Which was late 2022.
Yeah.
Right.
Kind of mind blowing to think that we're already three and a half years in.
Right.
But many of the things that people are talking about today are ideas that resurfaced or surfaced late 2022, early 2023.
And they've been given enough time to kind of unfold.
Maybe let me put it in this funny way.
Is anything that we've seen today in AI, which is amazing, mind-blowing, an obsession, is anything that we've seen totally unforeseeable given what you knew in 2023, early 2023?
I don't think so.
I don't think so either.
I don't think so.
It's not any differences of kind and it's only a matter of a difference of degree where the extent to which the speed and the extent is the surprise, but not the difference of kind.
And I think everything is basically a difference of degree.
Yeah.
That's true.
Actually, I'm in that camp too.
It's all just degrees cranking up and cranking and cranking until it just kind of looks a little bit different.
I'm with you.
That's like an abundant maximalist mindset.
I'm in there.
The birds are just dinosaurs with feathers, you know, whatever, right?
Like it's all the same thing.
Like it's just crank, crank it more.
It's on a continuum.
There's no one specific discrete point.
And so I think there are too many ideas in the graveyard of crypto that have been stymied by.
too many very intense blockers to conclude that the space is done.
In fact, some of the cranks have been really kept cranking.
Yes.
The infrastructural constraints, stablecoin popularity, regulatory adaptation, maybe a second new law.
We've already had one.
All those things, every time a new thing kind of unfolds, it...
affords an opportunity for a prior line of thinking to be resurrected.
That's absolutely right.
And that is kind of how I'm looking at it is like everything is either cranking forward or something's revitalized or whatever, whatever.
It should be self-evident that with the improvements in automation that come from AI, accelerating every aspect of maybe most tasks, right?
There's going to be more commerce.
There's going to be more management of capital.
There's going to be more intense demands on the allocation of capital and resources.
I could say a hundred things that are obvious entailments, but they're all going to touch finance.
If they touch finance, if they touch money, if they touch capital, then they're probably going to have to touch crypto.
And you could, of course, of course, of course, you could build this on a conventional system.
And I think a lot of people have a lot of anxieties about being re-intermediated by agents, right?
Basically, in one frame, you can disintermediate yourself from other sources, like whether it's like news or social networks or vendors, commerce.
You can say, I don't care about them.
I'm going to trust my agent who works for me to decide what to...
what to get from me or to be my intermediary, right?
So you are in some sense disintermediating yourself because now your representative works directly for you.
Yeah.
Instead of having to rely on the third party.
But you're also re-intermediating yourself because you used to have a direct relationship with the adversarial third party, maybe semi-adversarial third party.
And now you have this trusted agent that's in between you and the third thing.
So in some sense, it's intermediating.
In some sense, it's disintermediating.
And the right way to deal with this problem is to think like, okay, what are the ways I can constrain or I can observe or I can verify the behavior of the agent working on my behalf?
And that's financial constraints, informational constraints.
And as AI becomes more powerful, it becomes better able to circumvent constraints.
So how do we truly constrain it?
Cryptography is the answer.
I really think of cryptography for identity, cryptography for authentication, cryptography for money vis-a-vis blockchains, for verification of compute, for verification of programs.
There's so many layers that you're going to need to bring in to adequately constrain the AI.
So I really struggle to see a future where if you are preoccupied at all.
by the prospect of reintermediating yourself by trusting an agent, trusting an AI, where you wouldn't want to bring along a toolkit of things to alleviate some of the trust that you give to it in many dimensions.
So I see them as complementary.
Well, of course, we're at the exuberant adoption phase of AI where like...
Everything's good.
The AI companies are competing to give us more subsidized credits.
And like people are spending like 5,000 bucks a day on tokens, but only paying 200 bucks a month because like the companies are raising rapid capital.
Like to use the Chris Dixonism, right?
We're at the attract versus the extract.
We're in the attract phase, not the extract phase.
Okay, well, what happens during the extract phase?
Yeah.
That's exactly where crypto is going to come in.
And I think it's just a complete no-brainer.
I think it's a complete no-brainer.
And if anybody has a story for how that plays out without crypto, would love to hear it.
But I don't see it.
Yeah, I completely agree.
You've actually written a lot about this too already in some dimensions, like proof of personhood and various things.
But it plays out at every level.
We'll see which things work and which don't.
I don't know, right?
There will need to be constraints placed at different parts of the thing just to ensure that it performs what the user desires.
Yes.
And what serves the market best.
That's fantastic.
What's your favorite idea that was relegated to the dustbin that you think is ripe for resurrection?
I'm cheating a little bit in my answer, but...
The idea of the network token is the one.
It's not cheating.
It's the one for me.
The idea of an asset that is not a company but can capture value and can be community-owned and controlled without a single intermediary in the middle, a single human intermediary in the middle.
That's the most interesting one.
There's tons of other interesting ones, but that is the one.
And then other nearby ideas that I've always loved but are being developed, right, is...
The idea of a self-custodial bank.
It's pretty easy to custody your own stable coins.
It's a little harder, but still possible to add privacy to it, to the mix, through the use of certain protocols.
There's not that much more needed to go fully end-to-end and to really have a bank, so to speak, that is under your direct control.
I think that's really, really interesting.
And I think it's pretty obvious.
When you're self-custodial over your assets, and you have an AI agent that works for you, then the AI agent can control your assets.
And that becomes really interesting.
I mean, is it not self-evident that people will just trust the diligence and research of their gigabrain AI agent in the managing of their finances?
I think you've got to be insane to conclude that.
They won't.
I mean, apparently you don't believe that when it comes to music, but sure, I'm just kidding.
No, I agree with you.
No, it's true.
I mean, it's true.
It's true.
A weird interpretation of what Sonal's saying is like, you could almost have like an AI agent that like listens to your music for you.
Yes.
And just tells you, by the way, I love this album.
Don't worry about it.
You're busy working, but.
I liked it.
Does that relieve you at all?
That actually is kind of what I think.
I mean, isn't that what we do with our editorial when we just like distill people's voice into like a single Markdown file and then...
say, this is what you're using to then generate this copy.
I mean, it's not that different than that, except in your case, it's giving you that satisfaction, the psychological pill of, I don't need to go for completion.
My agent can provide the gathering and the completion, and then I can tune what to listen to, which could actually solve that.
Yeah, there's this Zizek anecdote about the Tibetan prayer wheel.
Okay, yeah, of course I know the Tibetan prayer wheel.
Yeah, so you put a prayer on the prayer wheel, and you crank it, and it's praying for you.
Basically, like because it's it's turning the scroll.
It's turning the scroll as if it was being read.
So it's like it's it's praying for you.
Well, if you just hook that thing up to a motor, you know, you can say a thousand prayers a second.
Yeah, just put it on a nuclear reactor or something.
Yeah, you'll get so many prayers per second.
You know, imagine how many albums I can enjoy as long as I just let the AI do it for me.
I mean, speaking of that, you mentioned having comparative advantages.
Is there a comparative advantage that you see humans having or what is one area where humans can remain?
Well, there's this great little tweet I saw by this guy who has this really cool AI-generated game in the browser about killing zombies.
He does a lot of AI game dev, and he just tries to do as much as he can with AI, basically, and test its limits.
And he highlights all these different areas.
He says, like, 2D art.
AI is basically at par with human artist.
3D.
human specialist, way better.
Game engine development, AI agent is better.
Story, story authoring, human, far, far better.
Game feel and like, you know, the sort of visceral experience of the game, human a thousand miles ahead, right?
So like, in other words, it's a spiky surface.
What human beings are better at versus what machines are better at, that's not something that we talk enough about at a wide level, like a wide cultural level, and it's changing too quickly, to have a grasp of like, this one thing is what the humans are better at.
This one thing is what the machines are better at.
That takes a while to develop the language around.
So the way to answer your question, Robert, is it seems like there's a lot of things.
Like right now, machines are arguably their best when it comes to coding a system that is according to spec and even brainstorming about the right spec.
They're very good at that, actually.
They're very good.
What they're maybe not as good at is like the feel of a thing, the coherence of a large interconnected world narratively or even like physically coherence at like a very large scale.
Some people like to say taste, which I'm like a little 50- I feel like you're contorting yourself not to use the word taste.
I knew you were going there and I was like, word is banned.
Well, I say that with a caveat because I think I obviously get what people are saying when they say that, but I wouldn't draw the line here.
I don't think the human being should die on that hill, right?
It's like the taste hill.
I don't know.
I've seen some pretty good stuff.
And I also have seen bad stuff with very terrible taste from plenty of humans.
Yeah, and the median human has horrible taste, by the way.
Yeah, exactly.
They're not all Anna Wintour.
No, exactly.
Why are we hanging our hat on that?
I don't think that's a solid thing to hang your hat on.
So to answer your question, Robert, I think that we don't really have a deep enough understanding or enough direct exposure and experience to have a crisp view of what the humans are better at and worse at.
But I think it will remain true that humans are going to be...
a little bit more adaptable, a little bit better at certain classes of things, maybe complex bundles of things.
And we'll find out what that is.
So just to wrap up, we've heard your story, your evolution.
You've been very humble, but very self-aware, which is really nice in not being annoyingly ponderous, but in the most interesting way.
What do you do for fun, Eddie?
What are your hobbies and interests that you just kind of do on the side?
My number one interest right now is definitely my little son.
He's so sweet.
He's so cute.
He's amazing.
I love him so much.
He's ridiculously cute.
Yeah, he is.
I do happen to find a little bit of time now and then to play video games with friends, like my old friends.
We've had like a group of ours Discord server for 10 years now, believe it or not.
Literally a 10-year-old Discord.
Video game world today is very interesting because we are really in the midst of a...
indie developer renaissance.
Yeah.
Because the cost of making very creative, weird games that would never have existed before has gone to zero and has been this way for a couple years now.
So I love trying like random, you know, little indie games.
Love that.
Play those with my friends.
We like what some people now call friend slop.
What's that?
I've never heard of that.
That's like co-op games that where you and your friends play together.
Peak is a great one.
There's a bunch of other great ones.
I'm really into shooting sporting clays.
Really into that lately.
My brother and my dad are coming up soon.
We're going to go shoot sporting clays again.
I used to have a lot of side projects at once, but honestly, like so much of my energy is now living out my fantasy of being able to build every software idea I've ever had.
Yeah.
Right?
I mean, it's kind of, it's honestly insane.
Is that really happening?
for you?
Yeah.
No, honestly.
Yeah.
I mean, I'm not going to, we're not going to turn this into an episode of me talking about like AI stuff, but you know, just like about like, like personally I stuff, but it's true.
It all works.
How much of your time is spent on like, you know, traditional corpo apps versus like any custom tailor-made bespoke homebrew apps?
It's honestly like 50-50.
Wow.
I spend a lot of time on goofy fun stuff.
I have a thing right now that like, I can give it a voice memo.
And it will make an app that deploys on my internal network and is just connected with all my other apps and my personal service.
It's so good.
Everybody has that.
That's like wow right now in like two months.
That'll be like the least interesting thing ever.
So yeah, but I love spending time on things like that.
Is there a holy grail of something you want to build that you're like, man, I just want to build this one thing?
What's just out of reach for you?
The truth is, I don't know if anything is.
I think it's just about, I think it's just having time.
All right, no more excuses then.
It's just having time.
There are no excuses.
It's just, it's all just prioritization.
Yeah.
I have like a server rack in my closet that has like, I think 200 terabytes of capacity.
Oh my God, that's like a heating disaster.
It does create a lot of heat, but like I do a lot of data hoarding.
Like books and, you know, music and stuff that I like to keep copies of because no one's going to take away all this beautiful stuff from me, right?
Yeah.
And that is so highly automated and so beautiful now.
Oh, wow.
It's so crazy.
It's crazy what you can do.
You can have a whole thing where every time you watch a video, you can have it downloaded automatically and stored forever.
I mean, you can have a thing where every time you listen to a song, a single song, it will go find the whole discography.
Yeah.
Of that artist and store it forever.
Yeah.
I don't know about you guys, but I feel that is where I feel overwhelmed.
And I miss having like scrapbooks and photo albums.
I feel like this digital abundance on this front is exhausting to me because I don't have the same curated photo albums I used to.
growing up.
And this includes all my other artifacts.
I do own physical copies of all my favorite movies and DVDs, but that's because I'm a bit of a prepper.
And I do think of like dystopian times when you might not have connection.
It's the same idea because the thought I've had, and it's kind of happening now is like, I've for years and years and years, I've store all this stuff.
Like I have, for example, I have every picture I've ever taken since 2003.
Since you had your phone.
Yes.
Yes.
Every picture I've ever taken.
How often do I go look at him?
Never.
Basically never.
But very soon, arguably today, you can ask an agent, like, make me an album of, like, all the most interesting pictures from that place I went summer 2005.
Yeah.
And it will do a really good job of that.
It will already do a pretty good job now, but probably just one more generation of the image recognition, image recognizing models.
And they'll be able to kind of like remember, oh, that's these two pictures are the same person.
And like I can intersect.
Yeah.
Kind of like it feels like it's all at the fingertips.
Another thing I'll throw out is like, I think everybody should lean into and enjoy the sensation of AI psychosis.
Right.
I think that's like a it's like a good thing.
OK.
Now, no, I'm not saying psychosis is a good thing.
OK.
People who have serious mental issues, that's not good.
You know, I hope they get help.
Right.
Like, but often when people refer to AI psychosis.
What they're actually talking about is a kind of gatekeeping.
Oh, interesting.
What it is is it's like, what is AI psychosis?
And it's like median, not in the extreme someone's actually having a problem, which by the way happens all the time.
People have- It happens in IRL even without AI.
But when you see kind of people throwing out in a flippant way, what are they talking about?
What they're saying is that the production of mediocre stuff at full speed without self-awareness is cringe.
Oh.
That's what they're saying.
They're saying it's cringy to not be aware.
You've become irrationally exuberant about your newfound powers.
That's what it is.
So when I say people should lean into the feeling, I don't mean do bad stuff.
What I'm saying is the people who are exhibiting some of the most intense public displays, it's that they just realize they can do stuff and they've lost their mind.
Like Superman.
Because like we were just saying, It's just kind of a matter of time.
The biggest constraint on me, no question, no question, no question, it's time.
On the weekends, during break, I produce the most stuff because I just finally have a little time.
You multiply time too.
Just to do things.
Yeah.
So what do you call it when someone is exuberant, producing maybe things that aren't to your personal standards, but they're incredibly happy and they feel that they now have tremendous unlocked.
creative potential.
I think that's what people are actually talking about when they talk about AI psychosis a lot of the time.
Of course, I'm not saying that everything produced is good, but I think we can all agree, anybody being reasonable should agree that AI is going to get better.
People's ability to control these things is going to improve.
And so what you're seeing is the beginning of a new, a revelation of a new kind of creative potential.
So I'm very pro AI psychosis.
I'll say that.
It kind of reminds me of, you know, things that have happened throughout history with communications technologies fading back to like, well, forever.
But, you know, to think about something like the Gutenberg printing press and him producing Bibles and wanting to get financial sponsorship from like the local bishops.
Yeah.
You know, thinking it was going to be this great thing for the church.
And then all of a sudden people got their hands on the technology and were making pamphlets and stuff in vernacular language and this and that.
Like everything was just exploding.
And then bringing that all the way up to like when people, when the internet went commercial and like all the forums opened up to everybody and it was eternal September.
Anyway, it's just like this thing happens all the time.
Like everybody gets access to the new thing and then they start doing all sorts of wild things.
Yeah, yeah.
This is, yeah, which I forgot to say earlier, but like a thing.
So here's like a medium-sized hill I'm willing to die on.
Oh yeah, you do get the answer to the smallest hill you'll die on.
Then I'll give the smallest hill.
So the medium-sized hill I'll die on is that most things, most people do most of the time.
is not worth much.
I mean, in the productive sense, I don't mean people's personal lives, obviously.
I mean, like, like the median website is pretty crappy, crappy.
Okay.
Sorry.
Like the median book is probably crappy.
The median article is crappy.
Like the median album is crappy.
Like, and I don't mean that to denigrate the creators of those things.
Everyone should endeavor.
And like, that's how you find the amazing stuff.
Right.
But like, there's a certain amount that has to be produced.
to make those gems, right?
Technology accelerates the production of mediocre things as a side effect of producing the incredible things.
Yes.
You can't throw the baby out with the bathwater.
So whenever you see something that increases that froth, that's good because that means there's gems coming.
And the only thing that's going to help you wade through the low quality medium thing is more technology.
Exactly.
Technology is a very solution to the technology it creates.
Exactly.
The sensation of something being bad is that nothing's happening.
The good sensation is tons of stuff is exploding, exploding out.
Just like it's ejecting tons of interesting material.
That's the good thing.
So that's my medium hill is like, your taste isn't that good.
Don't be judging stuff.
Let the market sort it out.
And then smallest hill is that ice cream and soft server should be maximum fat.
Oh, I completely agree with this, Hill.
Froyo should not exist.
Yeah, low fat.
Yeah, what about frozen yogurt?
Like it should not exist.
Low fat, throw it in the trash.
I don't care.
I hate it.
Maximum fat, 9%, 11%, 13%.
Give it all to me.
It's probably better for your health, too.
Of course it is.
On the protein and on the other side, too.
Of course it is, and I don't care.
Yeah, here we go.
That is a fantastic note to end on.
Eddie, you have been such a joy to chat with.
Thank you.
We love working with you.
I know we've all worked together now for a number of years, and we're so excited about you in this new role and kind of seeing what new philosophical degrees get unlocked.
But thank you for joining us on this episode of the A6NZ Crypto Show.