在加密市场运营一只10亿美元规模的市场中性基金是什么样

We spend a lot of time looking into risk because our concern was, well, what if DeFi is uninvestable now because Claude, Mistral, and Fable can just destroy everything? What our research has shown is that all the hacks—or 99% of hacks—that actually happened weren’t AI finding a way to break a smart contract. It was AI finding legacy OpSec vulnerabilities and exploiting them.

The metaphor I use is: imagine you have a bank whose security is up to standard. AI doesn’t invent an amazing drill that breaks through a metal vault; what AI does is find 100 old banks with wooden doors that you can knock out with a kick.

1. One Of The Last Market-Neutral Funds In Crypto

Welcome, everybody. Welcome to another 1000x show. We’re actually recording at the Out East Summit, hosted by The Tie. We’re super happy to be here because we have an amazing guest and a good lineup today. We’re in a beautiful vineyard. Welcome, Evgeny.

Pleasure to be here.

Welcome to the podcast. I want to start by saying you are one of the few market-neutral funds really left in crypto that are actually managing substantial capital. Maybe tell us a little bit about yourself first. How did you even get into this industry?

Yeah, we do carry the flag on the market-neutral side. I launched 3x5 five years ago. We started as a market-neutral fund, and that was inspired by my experience in TradFi. I spent 10 years working in that world—at first in hedge funds, in long/short, and then at UBS and Deutsche on the portfolio-management side.

Tell me a little bit about what it was like to be on an equity long/short desk back then. What were some of the things that you were looking at or doing? What strategies were you running while you were there?

That was a dream come true. It was a fundamental, discretionary strategy covering emerging markets. We had an opportunity to go into Egypt after the Arab Spring. We had an opportunity to go into some unique small-cap companies in Turkey, and to short the Chinese real-estate market by shorting iron-ore companies in Australia.

2. Shorting China, Trading Egypt After The Arab Spring

I think that inspired me to look for inspiration in unorthodox places, because that’s where it felt the value really lay.

Sure, actually, I want to go into that. You went into Egypt after the Arab Spring. Did you physically go into Egypt?

Egyptian stocks.

Okay. You went through Egyptian stocks?

Yeah. They closed the exchange when the Arab Spring started. We’d done a lot of work, and then we found there were some listed real-estate companies trading at 30 cents on the dollar, and we did some trades there.

Did you have to—how did you get the information needed to make those bets? If you’re going into Egypt, or Australia probably not so much, and all these different emerging markets, access to information is so important. How do you make sure that what you’re getting is real?

The beautiful thing about working in a large institutional hedge fund in TradFi is that you just get access to people through brokers. We would call a sell-side analyst and ask him to line up a few meetings with people who were running those companies. Back in the day, there were all these CEOs coming to sell-side conferences. They would fly to London, which is where I was working at the time, and then you’d go and meet everyone. You had access because you had capital.

When you say “back in the day,” is that no longer the case? You said CEOs were going to sell-side conferences. Now they’re not?

Compare it to crypto, where it’s an anonymous person on Twitter that you have to find a chance to speak to, rather than sitting down with the CEO.

Okay.

It’s a lot harder in crypto to get real, actual information out of people.

Legitimate information. Legit.

It’s very easy to have someone in a Telegram group or on Twitter, but actually to have an institutional-style conversation—that’s much, much less frequent.

But people still, in the traditional world, lie to you. I’m reminded of this story by one of my former bosses who used to work at Merrill Lynch. He was on the commodities-trading desk, and they invested in some coal mines in Indonesia. Evgeny Gokhberg

Then, 3 years later, it turned out those coal mines didn’t exist.

I’m wondering if you came across anything like that. If you’re going into Egypt, or Australia, or all these different emerging markets, access to information is so important. How do you make sure that what you’re getting is real?

You have to vet everything. There is no sacred place and no safe environment where you can operate. But there is a reason why governance is better in developed markets, maybe less developed in emerging markets, and basically doesn’t exist in crypto. These are spectrums.

The higher the level of confidence you have in information, the less you have to vet it. The higher the cost of lying is to someone, the more you can rely on it. But when you’re in emerging markets, you have to vet the information. Ultimately, you take a view.

Yeah. I’ve been in situations where there would be a CEO of a company, and he would walk me through a deck. Then I would ask him a couple of questions, like, “What if this forecast doesn’t materialize?” He would shut the deck and say, “Well, then we’re done.”

How do you even respond to that?

I’m like, “Well, this is very helpful. Thank you. I’m going to cross this off my list.” If the CEO throws a temper tantrum, I’m not investing in that company.

That’s usually a good sign.

Yeah. Okay, that’s good. You have to keep your cool.

I actually think we could do an entire podcast on the first part of the conversation. You could probably talk for 20 minutes about what happened in Egypt if you wanted to.

Yeah.

That would be fascinating.

Yeah, yeah, yeah. I started trading personally with tiny, tiny amounts around 2007, so I experienced the whole 2008 crisis. I wanted to be like George Soros and trade macro themes, so I spent most of my time looking at that when I was in the fund.

Chinese real estate captured my imagination, and I was looking into how to short it. You can’t really short anything in China, or at least you couldn’t back in the day, because it’s just not a market where you can actually do that. So we were looking for proxies.

I thought, “Okay, what will suffer the most if Chinese real estate starts to suffer?” What goes into building real estate? Copper, iron ore, whatever—or steel, to be more precise.

Then you look to see, okay, there’s coal and there’s iron ore. I don’t remember the numbers exactly, but commodities are extremely cyclical, just like crypto. When you have a commodity that’s in deficit, but the marginal buyer is starting to fade away, and these companies have a lot of leverage, you can have a one-way street.

We shorted some iron-ore companies, and that did pretty well.

That was a lot of fun. When you say that, did you have to specifically figure out which iron-ore companies were selling directly to China? When you construct a trade like that, what’s the process?

You work backwards, right? It’s like a domino that you have to connect backwards. What does China need the most? Steel. Great. They also need 10 other things, so which one is most vulnerable? The steel supply chain is most vulnerable. Great.

Then, within the supply chain, who do you have exposure to—people on the smelting side, on the distribution side, and so on and so forth? Then you have the suppliers of raw commodities into that. You look into coal, and it’s like, well, actually, there’s nothing clean to short. Okay, let’s look at iron ore.

You go all the way up to the underlying business and then find something that has a very clear correlation. If Chinese real estate is down, these guys will lose 90% of their net profit.

That’s so interesting. It’s something that I think about a lot these days because of how AI has made research so much easier to do.

3. How AI Is Distorting Research (And Who It Fools)

Back then, you actually had to spend a lot of time digging into things and figuring out the downstream effects. I think it’s distorting the market in some ways today because everyone will now just go on Claude and say, “Okay, if there’s a bottleneck in AI, or a bottleneck in data centers, what are the downstream companies that are actually supplying this? What’s needed to build out a data center? What’s needed to fabricate these chips?”

You’re easily pulling up these companies. For example, CCOI had a massive ripper because I think a bunch of people just saw it pop up super early on Claude and bought it. Then it just didn’t materialize at all and sold off 70% because nobody had actually done the real research.

There’s almost both more opportunity and less opportunity in this deep, deep type of downstream research.

I think that’s a great point. We always knew the cool things that were emerging. A year and a half ago, we rolled out a fully autonomous AI-agent kind of venture fund.

It's tokenized on Base. Whereas, like, you buy a token, it takes your money and goes and makes venture investments on-chain—liquid venture. Through that, we learned a lot about how to use agents for investment research because it's highly experimental. We didn't advertise it or raise anything for it; we just shipped it, and it's actually done pretty well.

But what that told us is: What does that mean, like, a fully autonomous agent? It's literally making all of it. There's no oversight at all on your part.

Not now. In the beginning, yes, but eventually—

Who's feeding it the information?

So we built that framework. Imagine you're hiring an analyst today, right? From your experience, you tell it where to look and how to think, but then, once it gets all the information, it decides on its own.

What framework are you using? Are you using Hermes or OpenAI?

Yeah, and that taught us that you can use agents for a lot of stuff, but actually, you can't replace a senior analyst. Right? In our world, it can replace a junior analyst because it can do repeatable tasks, but we don't feel it's at the point where we can actually replace an expert.

So I think your point is spot on, because, yes, it's much easier to do your research right now, but it will also give you a lot of false positives.

A ton.

Because if you're a person who doesn't know what you're doing—if I start reading about biotech stocks, I don't understand this sector. It will give me an answer, and I'll think that's the truth. Then, if you're a specialist, you can deduce what Claude is telling me, and you'll find alpha in the spread. So it's changing the whole game.

This is what people miss a lot of times about AI. They assume that its capabilities are that it will just give you the right answer. No, it's not. It's just not. It's not that at all.

There was this one thing that went super viral called World Monitor. Some guy rebuilt Palantir's dashboard for monitoring the world, and it went super viral on Twitter. The first question that you have to ask yourself is, “Okay, but is the data right?” I mean, the entire point of Palantir is that if you pay for their data, if you pay for their monitor, the actual numbers are correct. The actual things that you're looking at are correct.

Exactly.

And there's kind of no way to tell, right? The entire premise of the business is that they've spent billions and billions and billions of dollars on making sure that it's right. They've hired a ton of people. They have on-the-ground individuals detailing all this stuff. You can't just hope and dream that AI is going to do it, which is, again, why experts are still very valuable. But—

I agree. People have very short attention spans and want easy fixes to their problems. So they think one message to Claude will give them an answer. But that's not how it works.

But you can use it effectively, right? If you collate all the data, find all the different companies, and then you yourself have at least narrowed the space that you could search. I think that's one of the most valuable things, especially in investing: narrowing the space that you have to search for your answer. So I think it accelerates your work by a magnitude of 10x, if not more.

But it doesn't replace the need for you to think.

Yes.

And I think that's where people get it wrong, because they just stop thinking. They just ask the question and take the answer without actually processing the information themselves.

So you were in that world for almost a decade, or more than a decade.

About a decade.

About a decade. And then you find your way into crypto. Why did you leave? What drew you out of your seat, which presumably was actually quite interesting and fun?

I was working in a well-established investment world, and I was thinking, “How do I 1,000x my personal growth?” It became very clear that the hedge fund world and the banking world are too established and still restrictive, especially in London. Europe is probably a little bit different opportunity-set-wise, but coming out of the European Greek debt crisis, the opportunities were pretty limited.

4. Why He Left TradFi To "1000x" His Career

So I was thinking, “What is the most realistic way to meaningfully grow my career and opportunity set?” It was doing something exciting, and I couldn't find anything in the TradFi world. So I started educating myself on tech and fintech, and that's how I came across blockchain.

Like most people, I had heard about Bitcoin in probably 2011 or 2012, but I didn't really understand it. It didn't click with me. Then, seeing blockchain and discovering Vitalik and Ethereum in 2015, that was my aha moment coming into the space.

It's actually quite interesting that you had a method for approaching this, which I think is very different from a lot of other people. Sometimes, when I interview people—or even myself—I think I had more of your approach, where I was looking for an edge. A lot of people just sort of fell in love with crypto. It seems like you were searching for an opportunity. Can you walk me through that process? How did you approach searching for that opportunity?

Yeah, I was trying to be systematic because I'm always hopeful to fall in love with something, but it never happens—or it never happened—and therefore you have to make it happen for you, right?

Right.

So I was thinking, “What is the thing I know nothing about but it's very important?” It was tech because I come from a very traditional economics and finance background. So I convinced my boss to pay for me to do a course at Stanford. As I did that, I was like, “Okay, now I kind of understand tech. I kind of understand finance.”

So I started looking into fintech. And I was like, “Okay, I want to learn. What's the fastest way to learn? To lose money. What's the best way to lose money?” To be an angel investor.

Yeah.

So I made a few angel investments in the fintech world. Through that, I discovered blockchain.

Okay.

And I was like, “Okay, this is the whole crypto thing. It feels very scary. Lots of negative articles about it. Maybe it's a scam.” So I made the decision not to make any investments for the first year. I decided to allocate 2 evenings a week to going to meetups, lectures, and seminars. By virtue of doing that, I attended 100 events over the course of the year.

You're a busy guy. 100 events over the course of the year.

Yeah.

Well, if you think about it, 2 nights a week is not that much.

Yeah.

It's just all about consistency, and I think it's all about really wanting to get something out of it.

Then, through that, I was lucky enough to attend a meetup where Vitalik was speaking. That one meetup was my gateway into this.

What year was this? When did that happen?

2015.

2015. Okay. So you hear Vitalik speak and decide to yourself, “Screw ’em, screw TradFi, let me get in.” What was the first step that you took?

The first step was trying. I attended a lecture where Vitalik was speaking, and I didn't understand anything he was saying, but it was very clear that this was the future and that this was the beautiful mind building that future.

So I went very deep into the Ethereum world. Then I started proactively building a network, meeting people, and doing research in the space. Eventually, I ended up investing personally.

Okay. But now, a lot of people, I think, when they get into crypto, start off by just directional investing. Maybe you buy some Ethereum, you buy some Bitcoin, you buy this or that. What you do now is run systematic strategies to take advantage of the cryptocurrency markets and all the inefficiencies that exist. When did that start happening? When did you start building strategies to trade crypto?

When DeFi went live. Most people do come and buy coins, but most people don't have a system. They don't have a framework, right? They buy because a friend tells them. But if you buy something, you don't learn anything, right? So you need to choose where you want to learn.

When you're starting to learn, you want to understand if you can have a method and a sustainable process, because it's one thing to trade on the weekend. It's another thing to build a business, go out to people to raise capital, and then hire people whom you're responsible for.

When I saw DeFi, that was my second aha moment. Prior to that, it was very exciting to invest in venture and to invest on the liquid side. But this was basically buying a story and buying a promise of the future and the promise of a brighter tomorrow because there was nothing fundamental, right?

There was, “Oh, let's build this network, and then maybe people will come and use it. Let's build an Uber on-chain. Let's build an Airbnb on-chain. Let's build this and that.” But there was nothing fundamental.

5. The DeFi Aha Moment & How Yield Actually Works

When DeFi started in 2019, it became very clear that, for the first time in the history of crypto, we had real decentralized apps that you could use and that would monetize. They would charge you a fee in some shape or form. When you have an app that charges you a fee, whether it's on the App Store or on Ethereum, it now is a real business because you can value it, measure it, and monitor it. That gave me confidence to dive into it full-time and build a career around it.

Because that was an ecosystem where you could make liquid investments with a framework, and you could earn yield with a framework. I think DeFi yield has taken on a bit of a negative connotation these days because, when you look back at 2021, DeFi yield was effectively these companies issuing tokens to people for usage of the platform and basically paying far too much for customer acquisition.

Yield back then meant parking your capital in some pool using an application and then just getting paid out by DeFi protocols for doing that. When you talk about yields, I can't imagine that's what you're still doing. You must be talking about something slightly different, or is that still the world of DeFi today?

Right. So, I think that's a great question because it comes down to a framework. To me, DeFi yield is something very simple: I would speak to you and suggest that there's an opportunity for you to invest $1 of your capital, right? And you'd ask me, “Okay, that's great. What is the risk that I'm taking?”

As long as I'm delivering you a strategy where you only take smart-contract risk—the risk of a DeFi platform collapsing from a software failure rather than a market failure—that's what I classify as DeFi yield. And then there's the source of that deal, right? Is it what you said, which is basically companies using their equity as a marketing budget—“Hey, open an account with me, I'll give you a piece of my business”—which they did with tokens?

That obviously is dead because the token market is dead. But there are other things you can be doing within the space where you're getting paid by virtue of being an operator in that platform or in that ecosystem. And we have a whole framework that we've developed around that.

Can you talk to us about maybe a specific type of strategy that you're running right now that illustrates what that looks like concretely? I think sometimes it's hard to wrap your head around: If I have this dollar, what am I actually going to do with this dollar?

Yep. Actually, DeFi is way overcomplicated for what it is. These are very, very simple concepts, right? You get paid when you process someone's trade. So, you put money into a Uniswap pool, you take one stablecoin and another stablecoin, and then you collect a fee. Maybe that deal is attractive to you, or maybe not. If it's not, you go further.

Then you can lend, and you can lend your stablecoins against coins which are very volatile and very risky, or less volatile and less risky. More recently, you can be lending against RWAs—basically tokenized credit—which is great because we're basically back to what TradFi is.

But the thing that's been most exciting more recently is tokenized stocks on-chain, because your yield—or at least our yield—is an expression and a derivative of the market environment. When it's a bear market, there is less appetite to trade. There's less appetite to lever up, and therefore, if you're an economic agent that benefits from servicing the ecosystem from a liquidity standpoint, you make less.

When there's a bull market somewhere, you get paid much more because people are willing to pay more for leverage. They're willing to trade more, so on and so forth. There's a bull market in AI stocks right now. So, the same things we used to be doing with memecoins or other assets, being fully market-neutral, we're doing with tokenized NVIDIA and other assets like that.

And so, why is there a spread between whatever yield you can get doing this in the traditional markets and the yield that you could get in crypto? I think that's a big question a lot of people ask: Is there a yield in crypto specifically because there's a higher risk associated with that?

What you're getting compensated for is the risk that Claude models come in and break one of these platforms. We get compensated for the fact that whoever you're lending to is actually at a much higher risk of default just by virtue of being in crypto. So, maybe talk about that.

So, there are probably 2 answers to this question. On the AI side, I can address the second half, but I think people fundamentally misunderstand where the yield comes from, right? The yield doesn't come from someone coming to you and saying, “Hey, this is a risky platform. Therefore, you get paid.”

You actually get paid for being a lender. You get paid for acting as a market maker. These are the roles that, in the traditional world, belong to prime brokers, banks, and the Citadels of this world. In this world, they're outsourced.

My inspiration to start a firm wasn't that I was excited about a specific strategy. It was a thesis that DeFi is only different from TradFi for 1 reason, and that is that the role of a platform is separated from the role of liquidity providers.

So, if you have a bank in a traditional world, you walk into a building—it's a physical interface—you talk to someone, the bank matches borrowers and lenders, and the bank has its own capital. If you think of Uniswap, it has an interface and it matches buyers and sellers, but it has no capital.

The role of a capital provider is outsourced to the market, to an individual market participant, or to a professional market participant. So, we do what banks do in the real world. We provide liquidity to the ecosystem, and what a J.P. Morgan is earning because someone is trading or levering up on NVIDIA with them, we're doing that on-chain.

We're not getting paid because DeFi is risky. We're getting paid because there is economic activity on-chain, and then we're the ones making it happen. Avi Felman

And that's not something we're competing against at J.P. Morgan, right?

If someone wants to borrow against tokenized NVIDIA, we're explicitly competing with the traditional world.

These worlds are converging and colliding very quickly. We're speaking to banks that are looking to lend stablecoins into some of the vaults that we're running. We're looking for people to tokenize their equity or their credit and put it into the DeFi vaults that we're managing.

We're looking for people who come to us and want to take a loan on the back of a real-world position they have with a reinsurance or a credit investment.

Can you explain the concept of a vault to people? When you talk about DeFi vaults, I assume that half of our audience is familiar with it, but there's probably a substantial number of people out there going, “What are you talking about? What is a DeFi vault, and why are DeFi vaults useful?”

It's a great question. It goes back to my assumption and thesis that DeFi terminology is just terrible. The word “vault” doesn't mean anything.

People just randomly came up with the word DeFi one day, too.

And now we have to explain to normal people.

Yeah, exactly.

So, a vault is basically an investment box.

Yeah.

Right. Inside that box, you can put a memecoin. So, I can say, “I have a vault that will buy memecoins for you,” and you're like, “Well, great. Thank you very much.” Or I can have a vault which is running a hedge-fund strategy.

Right. So, just like the word “fund” doesn't mean anything, you need to understand what the fund actually does. The same goes for the vault. You can think of it as an automated financial activity. Sometimes it's all about finding a trading-strategy edge. Sometimes it's basically running a lending book, so it's acting as a risk manager in the bank, where you're deciding, “Am I giving out the mortgage to person A or person B?”

I guess the obvious follow-up question there is: What makes that more useful than just a fund? Why is that a better product to offer people?

It's not that I would think of it as a better product. It's a technical evolution of what the product looks like. For example, 30 years ago, if you wanted to invest into a fund, you would have to fill in a form. More recently, you can call your banker and they will buy the fund digitally for you. Or you can buy an ETF because it's much cheaper to launch ETFs these days.

It's simply a different wrapper. It doesn't make it bad or worse, because you can have an amazing fund which is extremely clunky, and you can have a terrible vault, or vice versa. It makes things more accessible. It makes them easier and more transparent.

But these are the features that can't really make a strategy better. They just make explaining and selling that strategy better.

That makes sense. And so, the types of people that are interested in investing—so, you run DeFi vaults, you run individual strategies. One question is: Who is actually putting money to work here? You ran over $1 billion. Whose money is that? Who's coming in? Who's interested? Is the market growing, or is the market shrinking?

In terms of who these people are, it's a broad mix. When it comes to our funds, we've been approved by several very large global investment banks to be put on their platforms.

So we have banks and clients of those banks. We have foundations, endowments, fund of funds, and professional investors. When it comes to DeFi vault market participants, most of them are on-chain users. Well, not most—all of them are on-chain users—and some of them are large crypto holders, like DAOs or whales, or just individual people depositing into vaults.

In terms of whether capital is growing one way or another, our business has grown quite meaningfully over the last 5 years, even though the space hasn't really grown that much.

And why do you think that is? Why have you grown when other people haven't? What are you doing differently?

I think just surviving. I think I ask myself that every day. If someone was early to crypto, they could have become early to DeFi, right? And then most people who were into crypto were reasonably young and started playing with it without a very clear strategy. It's quite rare, I think, for someone to have spent 10 years in TradFi and then go into DeFi.

So we have a 5-year track record in our flagship DeFi yield fund, going back to the early days of DeFi Summer. What gave us the confidence to launch it was the fact that we were in DeFi since day 1.

Right?

What also gave us confidence is that we've been able to develop a risk framework around DeFi, because you're right: people don't understand it, and they're very anxious about it. But when you actually develop a whole risk framework, then it's not rocket science and you can make it work. Not everyone was following a risk framework. Lots of people were looking at me like I was an idiot when I wasn't getting 20% yields in Anchor back in the day.

Right?

Some of those people are no longer around. I wish someone would tell me the secret sauce, but it's just sticking around.

It's just sticking around.

It's just showing up and executing.

You were running a fund during the collapse of Luna. I mean, that was probably a very interesting time for your fund. Did you ever have any exposure to them? You never did?

Yeah. We looked into Anchor, and it just didn't pass RDD. So we red-flagged it, and we have never had the same mistake twice.

Do you see any correlations between what Michael Saylor is doing and what happened with Luna, with Strategy?

Not really. Not at all, actually, because Luna was debt-backed by equity, and that equity was worth nothing because it wasn't doing anything. Here, we have basically a Lombard loan.

Right.

6. Surviving Luna, Saylor & The Real AI Risk To DeFi

He's holding an asset that's worth $100, and he's borrowed—I don't remember the latest numbers, but let's say $30. So could this be a situation where Bitcoin goes down by 90% and he could then be forced to sell, pushing the price down further? Yes, of course. But at the current levels of leverage, it doesn't feel like it's a risky thing.

Do you foresee yourself getting more involved in traditional markets now? I mean, people are talking about the fact that crypto—the economic activity has sort of taken a hit. We're in a slow period of the market, and the equity markets are actually trading in ways that they've never really traded before. There's so much activity happening there. Do you ever look over and say, “Hey, there are some strategies that we run that we could probably do in the traditional markets as well?”

There's always a temptation to go beyond your core area of expertise. So far, we've said no to that because we've built a very considerable, kind of superior edge around DeFi risk management, and that's our core expertise. When I think of myself going to our investors or potential new investors and telling them that I can trade Apple better than Citadel, I don't think I can do that.

As these worlds converge, though, there are going to be more opportunities for DeFi lenders, market makers, and traders. That's where we can benefit, because we can structure those trades, find spreads, and find arbitrage. But we have to do it on-chain, which is where our strength lies.

I'll also take a contrarian view on the crypto economy itself. We think, actually, it's likely to accelerate. We're more bullish on the market rather than bearish, and that's why we want to stick to our guns.

Do you guys do any market making on platforms like Liquid—or, sorry, like Hyperliquid or Lighter?

Some, yeah, some.

So when you talk about DeFi strategies, what right now is interesting to you in DeFi? I think a lot of people have sort of taken their foot off the gas. They're not really looking at DeFi anymore. Are there any innovations in the last 6 months, 12 months, or 18 months that excite you?

I think when it comes to what's exciting in DeFi yield right now, it's everything around tokenized stocks. The spreads you can earn and the trades you can do in a fully market-neutral fashion are interesting.

In terms of new opportunities in DeFi innovation, you get very jaded and very blurred at some point, because you would have seen 10,000 platforms. The coolest thing we've seen more recently is probably tranching. It's been around as a concept for a long time, but only recently have we seen people starting to execute on it.

That's basically rebuilding traditional finance. I wouldn't say it's very novel from a financial-engineering standpoint, but it's definitely novel to DeFi, and it's useful.

Do you need a really robust tech stack to be able to take advantage of these opportunities, or are there things out there that your average person could go in and allocate their capital to and earn a good risk-adjusted reward?

I think the hard part is not the tech stack to allocate; it's the ability to risk-manage. Because you're putting money into a DeFi platform, how do you know it's not going to get hacked? What's your framework? I think it's very hard to do that as an individual because the process is unclear.

We think of DeFi as high-yield loans, right? You have a set of platforms that sometimes collapse. So as long as your yield is higher than your cost of default, then you're okay. But as an average person, if you show me a couple of bonds issued by companies I've never heard of, how am I expected to spend my Saturday evening doing my research? You can't really do that, right?

Right?

So when people go into DeFi, realistically, they go blindly. They go on the back of a brand that they trust and a CEO who is active on Twitter, but they're not really doing their research.

Which sometimes is good and sometimes is bad.

My point exactly. So it's pretty dangerous, and you need to get lucky, right? You need to get lucky to not get wiped out, because once you do, you're gone, right? You're hoping to earn 10%, 20%, or 30%, whatever. But if you lose 100% in the second after you deploy, that's it.

What do you think is a reasonable yield right now in DeFi?

We're getting 11%–12% consistently. That's pretty solid.

When it comes to actual specific platforms, I'm curious if you have any takes on what platforms are actually in that high-yield bucket—or, sorry, in that high-grade bucket.

The way we operate is that we've developed a probabilistic risk-rating system for DeFi, and we rank everything as AAA, BBB, and CCC. Obviously, the usual suspects would be in the AAA category.

Are there any things that you specifically look at that might not be obvious to people in how to rank these in terms of safety? What goes into the safety rankings?

It's a very long process with dozens and dozens of checkpoints, but I would highlight one specificity that I think is absolutely critical in the context of AI risk. We've seen a record number of hacks in March and April. When you actually look at them in terms of dollar value lost, it's very normalized—it's below 2% annualized—but the number of events is very, very high.

So we spent a lot of time looking into risk, because our concern was, what if DeFi is uninvestable now because Claude, Mistral, and Fable can just destroy everything? What our research has shown is that all the hacks—or 99% of hacks that actually happened—weren't AI finding a way to break a smart contract. It was AI finding legacy OPSEC vulnerabilities and exploiting them.

I spend all of my time talking to TradFi, and the metaphor I use is: imagine you have a bank that has security up to standard—a great metal vault. AI doesn't invent an amazing drill that breaks through a metal vault. What AI does is find 100 old banks with wooden doors that you can knock in with a kick. So we spend a lot of time looking into OPSEC, and that's probably the most critical part right now.

Do you think that we're going to get a ton of upgrades in the future and, at some point, this is going to become even more secure? Okay, so that's your take: we're actually trending toward security, not away from security.

I think that's the trend we're seeing very clearly. Out of the roughly 60 hacks that happened this year, we didn't even know 50 of those 60 platforms, right? If our job is to know every single platform, the fact that we don't know some platforms probably means they're not very large and not very serious.

Makes sense. Yeah, I think this is a good place to wrap, but thank you so much for coming on to the 1000x podcast. This is a lot of fun, and it's good to have you on at some point.

My pleasure. Thank you.

Nothing said on the 1000x podcast is a recommendation to buy or sell any investments or products. This podcast is for informational purposes only, and the views expressed by anyone on the show are solely their opinions, not financial advice or necessarily the views of 1KX Media. Our hosts, guests, and the 1KX team may hold positions in the companies, funds, or projects discussed.