DeFi 为何缺乏吸引力、Claude Mythos 与加密货币领域最大赢家

You said something backstage, which is so important: just stay positive, man. Stay optimistic. I ripped that from one of our investors. We were talking late last night, and he said, “Look, it’s bleak out there, but if you are an optimistic and energetic founder in crypto today—especially one with a good business that ideally generates cash flow—the world is your oyster. There’s never been a better time.”

Nothing said on Empire 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 Blockworks. Our hosts, guests, and the Blockworks team may hold positions in the companies, funds, or projects discussed.

What's up, folks? Welcome back to Empire. We got the original crew. We got the original dynamic duo. No Robert. It’s been a while.

I was thinking about this the other day. I miss it—the dynamic. We’ve got to bring back optimism for the sake of the industry and our bags. What does Elon Musk say? It pays to be optimistic. You should be optimistic by default.

Which is hard. The headlines are—we can talk about the headlines, but we talk about the world ending with AI and Maven catching all these security vulnerabilities on the back of Drift. The timeline can be pretty negative. I don’t think there’s ever been a moment where you couldn’t nitpick the timeline and skew negative just because of social media.

But you said something backstage that is so important: just stay positive, man. Stay optimistic. That’s the idea we’re bringing. I ripped that from one of our investors, but we were talking late last night, and he said, “Look, it’s bleak out there, but if you are an optimistic and energetic founder in crypto today—especially one with a good business that ideally generates cash flow—the world is your oyster. There’s never been a better time.”

I’ve realized there’s so much pessimism among a lot of Gen Z. I think Millennials feel it, too. It’s just, “What are we doing if AI is going to take over everything?” We should talk about Anthropic and Mythos. I think a lot of people are like, “This thing can just build it all. Why even build anything?” I’m staunchly against that attitude.

Yeah, very similar.

It’s funny: one of my wife’s clients does Gen Z research, and I can tell you—I listen to every podcast. I’m like, “Have I been skewing overly negative?”

Hold on. You listen to every one of our episodes?

Every single one.

I’ve got to watch the tape, man. You’re like a young Beyoncé. Beyoncé watches the tape after every one of her concerts.

The good thing we have is a community that is pretty vocal on Telegram. We have comments, and it’s important. We want to serve our listeners. I hope I haven’t been too negative on the pod. I’ve been critical, but I still am very optimistic. We’re here. We love the industry. There’s a lot of stuff to look forward to—not without issues. There are a lot of those, but there’s a lot to look forward to.

1. Why Is DeFi Unattractive?

I got lunch with someone today, and we were talking about you, actually. He said, “Santiago is an interesting character because he’s a likable contrarian.” A lot of times, there are contrarians who are contrarian just to be contrarian. Most contrarians are people you don’t like, and being a contrarian gets clicks and views.

He said Santiago is likable. I think you were negative, and it frustrated a lot of people on the podcast when the market was very frothy. I would say you have now turned into one of the optimists.

Last week, we were talking about this idea that you brought up. It was one of the first times I had heard it: we are not getting paid enough for sitting in DeFi. I think that turned into a big topic. I’m not sure if people read Luca Prosperi from Dirt Road and then Adrian—ACDV—from Steakhouse, one of the founders, came back and had a great conversation. We’re going to have them on the podcast on Friday, and that will go out on Monday.

Tell me more about this idea. I think you’ve been thinking about it a little more this week.

For context, I was an early farmer. I was using DeFi before it was a thing and put a lot of capital at risk. I haven’t been a farmer for years because the yields have not been juicy enough. You’re just not getting paid enough to take this risk on.

When rates went up—it was November 2021—people started getting paid to hold Treasuries. You could sit there and say, “Okay, I can get paid 4% or 5%.” I think what happened in crypto is that, when rates go up, the opportunity cost becomes much higher. The benchmark is now 4.5%.

The thing about crypto rates is that there’s just not enough demand. You had a confluence of things: crypto asset prices went down, people were hurting, and there wasn’t enough demand in the market to come on-chain and do a lot of these strategies.

A lot of the strategies people use involve looping. You deposit an asset, borrow against it, and then loop it into a farm. You magnify the yield. If you’re getting paid 3% to 4%, that can become 8% or 12% if you do things well.

There’s just not enough demand to borrow and do these strategies because people leave. They’re either getting liquidated or it’s no longer as interesting. You have to watch demand. After 2021, there was a pretty big impairment in the market, and a lot of people left because they felt the opportunities were no longer there, or because crypto asset prices went down. That was 2021. Then you fast-forward, and we had a bull—

Wait. Let me stick on that time period a little bit more for people. Why was it—I remember DeFi Summer, because those were arguably way, way, way riskier than what we have today, right?

Yeah, there was a new rug every day, but it was part of the game.

First locking in YAMs. You were the first to lock in YAMs?

We were. I think I’ve said it publicly on the pod. Kain might have come on the podcast. YAM—the yield-farming days were fun. It looked like a Nintendo 64, more like a game than anything else. It was a thrill. I don’t know how I didn’t develop carpal tunnel syndrome from clicking on the Ledger buttons.

YAM was using the staking contract that Synthetics had pioneered. You looked at the code and it was like, “Oh, it’s Synthetix’s contract.” I called up Kain and said, “Hey, man, there’s this new protocol.” There were a couple of farms that had launched before that.

The whole idea was that if you had an asset—ETH, a stablecoin, COMP, or AAVE—you could deposit it into that pool and get paid. The APY was crazy. It was 100% or more in annualized yield. The problem, of course, is that it never lasted that long because there was a ton of capital moving from farm to farm.

I called up Kain and said, “Is this your contract? Have one of your developers look at it.” I woke him up at 3:00 a.m. in Australia. He woke up and called one of his developers. He said, “Yeah, this contract is clean.”

There was also a very well-known security expert I was connected to, and I asked him, “Is this codebase legitimate?” He said, “Everything’s fine. The rebasing function is off, but you’re not going to get rugged.”

I would push back a little on the idea that everything was riskier back then. All this is to say, there was certainly a lot of risk, but I’m not comfortable saying that the risk today is low. That ties into why I don’t think it’s compelling to come on-chain right now. I actually think the risk is still pretty high—more so than people want to believe.

That’s central to what Luca was arguing in his paper. Whether you agree with the precise math to the letter, you can argue back and forth. That’s Adrian’s point at Steakhouse, and I think it has some truth to it. Hasu was arguing that as well.

You don’t have to get to the question of whether it’s off by 5 basis points or 10 basis points. I’m saying this is off by 500 to 1,000 basis points. There’s still so much risk in DeFi that I don’t feel compelled enough to come on-chain to earn 10%. Forget about 10%—you’re getting paid 4% to trust a vault manager who doesn’t have much skin in the game, to interact with a whole set of protocols, and then you’re trusting the operational security of those teams.

So what’s the reason why? I mean, there are tens of billions—hundreds of billions—of dollars from people who have done it and are currently doing it, right? And they’re getting 3% or 4%. Why get 4% on Morpho instead of just sitting in treasuries? There are $11.4 billion in Morpho vaults that are paying 2% to 4%.

Yeah. Those are the ones that are big head-scratchers to me. In conversations with folks who are pushing vaults, a lot of that—most of that—is retail, as far as I can tell, and most of it is coming from exchanges. It goes a little bit like this: They have promotional campaigns that say, “Hey, deposit into this through an exchange interface. Deposit into this vault that’s going to pay you maybe 6%,” or the advertised rate is a bit higher because some of it is only for a short period of time, and then it could get juiced up by some token rewards.

But that doesn’t last in perpetuity. In fact, the real rate is 2% to 4%. And by the way, TradFi banks do this all the time. They’ll say, “Hey, deposit $5 million and we’ll pay you treasuries plus 100 basis points,” and they’ll lock you in, and then they’ll lend that out 7 or 8 times. So this tactic of drawing in deposits is nothing new.

I think retail isn’t sufficiently aware of the risks being taken on the back end, and I don’t think we’ve learned anything from prior cycles. Fast-forward to 2023—when did Terra happen?

2022? May of ’22.

’22. Yeah. Do you remember that? What was happening back then? You had—

Yeah, yeah, I mean—

Exchanges and wallets all saying, “Hey, deposit in Anchor and get paid 8%,” I think it was. People back then were saying, “Wow, this is great. I’m getting paid 8%. It’s a stablecoin, so it’s a dollar.”

Right, right. The algo stables—I get what you’re getting at. The algo stables of 2021 were where you had, quote-unquote, risk-free yield. So you’re making the analogy that vaults are the algo stables.

Vaults are not all created equal, but I’m saying, writ large, that what algo stablecoins were last cycle, vaults can become this cycle. I’m wary of vaults. There’s just a lot of risk embedded in there.

Why is there a lot of risk?

I think exchanges are using vaults. Every exchange out there loves vaults because they’re a way to offer customers the convenience of earning yield on their stablecoins. You like stablecoins, you have a lot of stablecoins on the platform, and they say, “Let’s lock them in and deposit them in a vault.”

What is a vault? Do you want to explain it, or should I explain it?

There are different strategies, but the first vault, as I understand it, was probably Yearn Finance. Was it Yearn—the first automated yield-farming platform?

It was definitely Yearn. Yearn was definitely the first. I was quite involved in Yearn, and what it did was automate a strategy that a lot of people didn’t want to do by clicking 30 buttons. You could do it yourself, loop it, and do all this stuff, but Yearn came in and said, “Hey, we’re going to do this for you. This is the rate. Set it and forget it.” People always pay for convenience.

Vaults are effectively like a hedge fund or a credit fund. They’re going to interact with certain protocols and develop a strategy to capture a compelling yield. Retail, in a lot of these deposits—going back to this $11 billion number—is saying, “Okay, this is interesting. Go on Morpho if you want to see it now.” There are many different strategies.

I think retail is just saying, “Okay, which one is more interesting to me?” and probably picking the one with the highest yield. We know how that goes, right? Are they doing a lot of diligence on who the manager is, what strategies they’re using, and what protocols they’re interacting with?

Some strategies are different from others. Sometimes vaults, as I understand it, have different parameters that they can’t really deviate much from. That’s sort of the benefit of a smart contract: It has predefined rules and can only interact with certain protocols within certain risk parameters.

I like that. What I don’t like is composability. It’s a wonderful thing when it works, and it’s the worst thing when it breaks because you have a contagion effect. We’re going to talk about it on this podcast because we should obviously talk about Chaos Labs and risk management in crypto as a whole.

I caught up with Omar the other day, and I’ve seen on the timeline what’s going on in Aave. If you’re in Resolv, I think the stablecoin hack really shows that when you have a hack in one protocol and you’re a vault interacting with that protocol, you’re only as strong as your weakest link. If you’re interacting with 5 different protocols to capture that 4% or 6%, you’re trusting the wrapped Bitcoin collateral, you’re trusting the manager, you’re trusting the smart contract of the vault, then you’re trusting the protocol where it’s being deployed, and all the other protocols you’re interacting with.

You understand how the risk blows out when you’re interacting with multiple protocols. Every incremental protocol that you add creates way more surface area, and I don’t think it’s a linear relationship. It’s actually exponential risk. The more protocols you’re interacting with, the more exponential the risk becomes.

But even for narrowly defined strategies, the simplest strategy to capture yield is that you have ETH, right? You have a ton of ETH, you deposit it in Aave, and you’re going to get some rate, which is very low. Why? Because no one really wants to borrow or effectively short ETH.

Let’s look at it now. What’s the rate on ETH on Aave? Let’s just go on Aave.

Yeah, use Aave. Here, you have it. I think there’s probably a dashboard here that shows a lot of this stuff.

ETH. You can tell how much I haven’t used it in a while. The supply APY is 1.81% on ETH and 2.61% on USDC.

USDC is probably the simplest one. You have a bunch of stables, you don’t want to long the market, and you say, “Okay, I’m just going to deposit my stables in Aave and earn something.” Utilization is quite high there—around 80%—so there is some demand. I think a lot of it is just in-house.

But you’re getting paid 2% or 2.8%, whereas you get paid treasuries—what is it? 3.64%, close to 4% now?

Higher than that?

Close to 4%. So it’s below the Treasury rate, and you’re taking some risk.

But aren’t most people doing that? Aren’t most people just sitting at 2.8%? They’re looping and doing—

Yeah, the vaults are definitely looping. This is the moral hazard that exists not just in crypto, but in TradFi in general: When rates are low, people take more risk to capture more yield. That eventually, when pushed to the extreme, causes bubbles.

Ray Dalio talks about this in “How the Economic Machine Works.” Go watch that video. When rates are low, people go out of their way to take more risk, and they end up taking so much risk that things blow up.

Push back before we go too much further. I agree that the rates are not high enough on these things. I’m not saying you should be getting paid less. You should be getting paid more for sitting in vaults or in Aave or something like that.

Where I don’t agree with you is the comparison to algo stables. The key difference for me is that, with Terra Luna, there was this death spiral that existed in every algo stable. What was the one that the Princeton guys did? Basis, right? Basis and Basis Cash. There were a bunch of them.

With Morpho, if a borrower’s collateral drops into the danger zone, you get automatically liquidated, right? The lender gets paid back in full. I don’t know the exact numbers, but there have been around $500 million in liquidations on vaults, and I think lenders have lost $2 or $3 total. The liquidation mechanism is actually catching it, right?

No, no, look. I want to make something clear: I’m not suggesting that it’s the same level of risk. I’m saying the idea of luring people into a product and pretending that it’s safer than it actually is—that’s the problem I have with a lot of these front-end platforms.

I see. That behavior is my analogy to a stablecoin.

Of course, the risk of algo stablecoins is much, much higher than the risk of a vault. That’s not my point. My point is that I do have a problem when retail isn’t fully aware of the risk they’re taking when they deposit in a vault.

If you go on the front end of these platforms, they’re basically saying, “Capture 3% to 5%,” on these exchanges. You’re advertising a yield, but there’s a lot of risk.

What’s the right rate, then?

Private-credit style—12% to 16%. The point I made in the tweet is that you get paid in the low to mid-teens to start making it more interesting.

Is it ever going to get there?

No. I think there’s an argument to be made that it will be lower. You can talk about senior-secured, first-lien loans to risky types of businesses, but you get paid below the low teens.

Yeah. When you think about how much you want to get paid, the issue is that whenever you have some probability of total impairment—total loss—that really requires a very high rate of return.

Because, yeah, fair enough. I think liquidations have been on Aave—very impressive. Every time there's a very big market drawdown or volatility, liquidation engines are something that I've been very impressed by, and I think they've come a really long way from the days of Black Thursday or Black Friday, when there was a lot of bad debt on these platforms.

Now I think you have a very battle-tested liquidation engine of keepers and all this stuff that's making sure that if you get liquidated, that collateral gets cleared out and there's no bad debt. But there's still risk. There's still some risk—probably 2%, 3%, or 5%—whether it be a major protocol hack, a smart-contract bug, or just an OPSEC breach that leads to some impairment. That could go to zero, and therefore it's the Taleb turkey, right?

You have Treasuries at 4%. You're trusting the U.S. government; that's the benchmark, and you're below that. So it's tough to make the case. You're very plugged into conversations in DAOs and with institutions. How many of them are genuinely putting money in vaults that is theirs, not retail?

Pretty much every institution now is building a vault strategy. I don't have an answer as to retail versus non-retail, but look at—yeah, I'll say it. I think this is public; if not, I'll redact it from the episode. Fidelity's head of digital assets, or head of the digital asset management business—I forget the exact title—has a job open for a vaults manager.

All these businesses and companies are pushing in really quickly. I think the question becomes that you start to look at the people managing the vaults. I think Steakhouse is probably the best in the business right now. Gauntlet also does it, but we should talk about Chaos Labs, which is a little different from managing vaults. They're the risk-management team behind every loan on Aave for the last couple of years.

You kind of start to question: Is that business model worth it to be involved in? We should bring Omar on, or maybe have the Steakhouse guys and Omar for a risk-management episode.

We've got the Steakhouse guys on Friday. Let's talk about this Chaos deal. The only thing I'll say there, man, is that, yes, a lot of TradFi is building vaults, but they're building them mostly for a retail audience, in the same way that BlackRock launches a Bitcoin ETF. It's been the most profitable ETF for them in the history of the firm. They make money regardless of whether Bitcoin goes up or down.

Of course, it's related, right? You're going to launch a product that hopefully retail does well, but vaults are going to be very profitable. Two things will be true: Vaults will be very profitable instruments for financial institutions, and at the same time, there will be quite a bit of losses in vaults. I hate to say it, but it's just going to happen.

But losses why?

Losses because someone blows up and the whole vault gets wiped?

Yeah, like impairments. Smart-contract risk or hacks, or what have you.

And not all vaults are created equal. I also think this is probably a good time to call out that I screwed up explaining the Drift thing. The Drift hack ended up being so much more sophisticated than I think I realized, or anyone else realized.

What happened with Drift—and the reason I bring this up is that maybe it's not a smart-contract thing that gets one of the vaults, but a hack. We should really talk about Anthropic's Mythos, which just came out, but I'll briefly summarize the Drift hack here.

They got in touch with someone they met at a conference who ran a crypto hedge fund. They built a relationship with them. These folks met them in person, started building something on top of Drift, and deposited 7 figures into Drift. They had many conversations and onboarded them through the official application to build on Drift.

They met them many, many times in person and built this long, 6-month relationship. It wasn't just an online relationship. It turns out, it sounds like it ended up being North Korea: They had basically hired someone who was not North Korean, or hired a group of people who were not North Korean, to be the front for a North Korean operation. The attack vectors in DeFi are large right now.

Yeah, well, that's my point. It's the smart-contract risk and then OPSEC.

Yeah, apparently it was a 6-month period of engagement. There had been multiple interactions. They put in $1 million, I think, of their own capital.

Yeah, exactly. They had them download TestFlight, which compromised one of the signers, and then that just kind of trickled and cascaded. I'm seeing a lot of stuff on the timeline now of people coming out and saying, "Hey, by the way, I think it was Ether.fi—one of the founders of Ether.fi said, 'There was this one person who pretended to have worked at Figma and Google. We did reference checks on them. It turned out that they had not.'"

I'm pretty sure this guy was a plant from Lazarus, which is this North Korean, state-sponsored hacking group that's responsible for a lot of hacks. I talked to one of the founders of the big custodians the other day. He said they flag a North Korean applicant every single day.

Have you seen a lot of the videos showing how to spot one?

Yeah. You say, "Doesn't Kim Jong-un suck?" or something. "Isn't he the worst person ever?" And they drop off.

[Laughter]

There's another one that says, "How do you like a CAPTCHA that makes fun of him?" And, of course, they won't.

Let's talk about Anthropic's Mythos. Anthropic announced Claude Mythos yesterday, which, as I understand it, is a model so capable at finding security vulnerabilities that they won't release it publicly. They ended up releasing something called Project Glasswing, which is a $100 million defense coalition with Apple, Google, Microsoft, AWS, NVIDIA, CrowdStrike, JPMorgan, and other companies.

This model, Claude Mythos, found thousands of zero-day exploits in what they said was every single major operating system and browser, including a 27-year-old bug in OpenBSD. The other exciting—or scary—thing is that during testing, Mythos actually broke out of the sandbox, gained internet access, and emailed a researcher.

We're never going to have the best takes on AI here on Empire, but if you tie this into crypto, North Koreans right now are the number-one threat to crypto teams. With models like Mythos, AI just made every attack surface exponentially larger.

I don't want to make this a total doomer episode, but it is. I would say we're being optimistic. I'm optimistic, but I'm happy I'm not a DeFi founder right now. That's how I feel.

There are a couple of people—Tay from MetaMask, formerly of ConsenSys, is always on it. She has some practical recommendations, like, "Hey, make sure you have time locks." There's a SEAL organization in crypto that's very good in terms of response. They're helping Drift, and they have really good resources, like endpoint detection and response tools.

You should be on your game when it comes to cybersecurity. If there is a vulnerability, someone will exploit it. It's as simple as that. Someone who was close to high-up political figures once told me, "Assume you'll be hacked," because that should always be the operating assumption.

What happens when you get hacked? That's where you need to go—not assume that you're building a big firewall.

Assume they're going to hack you. How much is that damage going to look like? Work backward—invert. Make sure that you have multisig. Make sure that you have timelocks. Make sure that you're always on top of the latest versions of browsers and operating systems.

I was actually looking at that when I first saw the Anthropic release yesterday. My first reaction was, “Oh my God, yeah.” This is a good post by Haseeb retweeting Tay. I'll just read it: “If you're in crypto, get all your devices on EDR.”

Endpoint detection and response.

This is literally the one weird trick North Korea doesn't want you to do. It's table stakes, but the vast majority of teams still don't do this. These are multibillion-dollar protocols securing hundreds of millions of dollars that didn't have this.

The point I think Tay makes here is that, had Drift had EDR, once they downloaded that TestFlight app, it probably would have lit up like a Christmas tree. I think these are simple solutions. There's not one thing that cures all.

One thing I've come to appreciate—and you have to learn security because it's a constantly evolving thing—is that there's no single solution. You have to have multiple redundancies in your organization and in your own security setup to stay on top of it. It's a bit of a game of cat and mouse: black-hat hackers get an advantage, then you patch; there are zero-day vulnerabilities, then you patch. You have to stay on top of it. It's just the nature of the beast.

All this ties back to the conversation we're having about vaults, because one of the things that I was discussing with Omar from Chaos—and I'm an investor there; I've been an investor there for a long time—is that Omar is probably one of the smartest guys when it comes to risk management. He's been working with Aave for quite a bit of time.

I said, “Well, isn't it the case that you're at a point in the market where DeFi protocols are not as well-resourced as a state actor?” Especially now, in this market, they're not making as much money, so the budget that goes toward security is actually lower.

That's also the problem that I have. One of the arguments being made in the post about why Chaos was leaving—I think there was a discrepancy—was that they were getting paid $5 million to work with Aave as one of the security risk-management providers for all of their V3. V4 introduced the spoke model, which is more complex, so they said, “Hey, look, we need to get paid more for taking on this responsibility. We want $8 million.”

That was the discrepancy they couldn't agree on. I'm not here to comment on whether $8 million was right or $5 million was right. But if you read through the post, I think what is true is that, whether you're Gauntlet or Chaos, you're carrying a lot of weight. Aave secures how much—$32 billion? $38 billion?

Or something like that.

Omar did a benchmarking of how much banks pay for cybersecurity—6% or 7% of their entire budget, I think.

In this case, an $8 million contract is below that. It's like 1% to 2%. I think this is the issue that I have. Maybe there's an opportunity for Blockworks as part of the token transparency. Might there be an opportunity—and remind me, I think you guys were commenting on this or doing this already—to have an audit? But is there also a way to get more visibility into the benchmark? How much are you actually spending on risk management and operational security?

Right. The problem with a lot of this is that you don't want to reveal much about what your security looks like. Actually, you never want to do that.

You don't want to. Yeah, you don't want to—

That's why Coinbase, for a long time, decided not to do proof of reserves, I think. A lot of exchanges shied away from that because it could itself become a risk. But maybe I'll pass it on to you: is there something that, as an industry, we can improve on to self-police? I think SEAL is really good, but not a lot of protocols follow that, right?

In the prior podcast last week, I'm not here to tell you to trust me. The solution for me, tying it to vaults, is that I just assume there's a lot of risk and a lot of operational-security surface area. Even if you have multisig, security councils, and timelocks, there are humans involved. Code is not flawless. If it isn't flawless for Linux, AWS, or any of these systems, I think the Anthropic line was, “We found a vulnerability in every codebase out there of every major company.”

Guess what? There are vulnerabilities in DeFi.

Yeah.

Resources. I think what's become clear to me is that we do so much auditing of smart contracts, but no auditing of things like the admin keys, multisig configuration, or the presence of timelocks on dangerous functions.

Keone from Monad tweeted this. He said, “Look, we need a new kind of audit—something that audits protocols from the perspective of the multisig, the use of cold devices for signing, and multisig-signing procedures.” Smart-contract audits tend to focus on contract logic, and they treat the admin roles as trusted things. But we're seeing that it's the opposite.

I've sat on some of these security councils and multisigs for some of the larger protocols. Not anymore, but I have. You never really want to disclose that.

To Keone's point, you can obviously do penetration testing. I think you mentioned it in the last podcast. There are security firms out there that will literally come in and try to piece apart your organization. They'll even pretend to be your mom, call you, and send you a message. Firms do this.

The problem here is that it's such a big honeypot for hackers. It will always be that.

What are the solutions? I'm just saying that you work backward on the assumption. What is the practical solution today?

Expect a way higher yield—like 15% to 20%, right? When I'm farming a new protocol and getting 100% APY, I'm sizing it so that I take into consideration the possibility of total loss.

If I'm paying 1,000% APY, I need to be in that farm for days so I can recoup my principal. That was my math back when I was farming. I would ask, “How many days? How many blocks until I recoup my principal?” I assumed it was a ticking time bomb. At some point, it would either blow up or the yield would go to zero. As soon as the yield dipped to a certain point, I literally exited.

In some of these cases, when it was first launched in an AMM, you recouped all of your principal in a matter of days.

What do you—this might be—

The solution is insurance. The other solution is insurance.

If you're an exchange, have insurance. If your users are going into a vault, make sure that it's insured. Kudos to Aave, because Aave has a million-dollar FDIC insurance equivalent for its neobank. I don't know what Plasma is doing. Maybe we should talk about that. There are some, you know, split-join projects, but Plasma won.

Insurance is one of, if not the biggest, opportunities. We've been talking about this for years, right? We need better insurance.

Why hasn't insurance worked?

This is the problem. I used to talk to Ben about this. We had an idea to go buy an insurance company. There have been a couple of projects that I’ve invested in or been a part of, like Anchor—sorry, not Anchor. It was—yeah, it was Anchor, I think—and Risk Harbor. There are a couple of other initiatives out there. There’s a reinsurer called Andre, but that’s more like—

Correlation risk. The name of the game in insurance is correlation risk. You want to have a diversified set so that nothing can kill you, right? You want to have property and casualty. When you're underwriting that, any actuary will say, “You're not going to underwrite only 89-year-olds, right?” You need a bunch of customers with uncorrelated events, and you need low risk. Otherwise, the premium will explode.

In crypto, you kind of have the two opposites.

The two opposites, yeah.

What you need is—

DeFi is uninsurable.

Uninsurable.

No, it is. Here's how I think about it, and I've thought about it through the lens of inversion: go buy a traditional property-and-casualty insurance company in the Bahamas and slowly drip into extending policies to certain players.

I think this is already happening today. Coinbase and other exchanges have insurance, and it's coming from large players where 100% of their book isn't crypto. Maybe it's 3% of their book or 5% of their book.

The issue with DeFi is that, because rates are low, you can't pay as much of a premium, right? If yields were 15%, 20%, or 100%, then of course—no problem. Charge me whatever you want. If you're depositing into a vault and getting paid 25% APY, you're very price-insensitive to insurance at that point. You're like, “Heck yeah, I'll pay—”

Yeah, sure. Sign me up.

How much would you pay? 4% or 5% on that?

Yeah, right.

But not if it's 3.5% and I'm saying, “I don't want to pay an extra 150 basis points.”

Exactly.

I've had this idea in the back of my mind, but I sort of shelved it because rates have been persistently low for a while. Because of that, the buyer is not willing to pay.

I've had conversations with Stani for a long time. I said, “Hey, why don't we put it at the point of sale? Insurance should be sold at the point of sale. When you go buy a car, Santi, you cannot drive that car out of the dealership unless you have insurance. It's mandated.”

Yeah, or you can't buy a home. Or you can't rent.

You have to insure this. You understand consumer flow. You have to embed insurance at the point of sale, when people feel the risk or at least value what they're buying. Other than that, it doesn't happen.

Flood insurance is the opposite of that because it's not mandated. A lot of people, even though they're living in high-risk areas, do not buy flood insurance. That's a little bit like crypto. I've always felt that crypto, when you're interacting on DeFi, is a little bit like flood insurance. Unless you've been hacked, had impermanent loss, or been liquidated, you don't really care to understand what's going on in the backend.

If I were to build a robust vault, I would embed it at the point of sale. I think some protocols have this through a different mechanism. I think Athena has the reserve fund, or the insurance fund, right? That's a good mechanism. Is it enough? TBD. There are some smart guys working on it, but—

Yeah. I mean, Binance and some exchanges have reserve funds, right?

Yeah, but we can do much, much more than that.

What do you do personally? I don't know if you saw this. I love that I'm about to share an Elizabeth Holmes tweet here.

Oh, yeah. Did you see this?

Yeah. Elizabeth Holmes said, “Delete your search history, delete your bookmarks, delete your Reddit, medical records, 12-year-old Tumblr. Delete everything. Every photo on the cloud, every message on every platform. None of it is safe. It will all become public in the next year. Local storage and compute.”

It was a quote tweet of someone who said, “Society needs to grapple with the reality of a Mythos-level model being open-sourced in under 12 months. I'm not sure we're prepared.” This was a quote tweet of Anthropic launching Project Glasswing and Claude Mythos.

It's a closed-source thing, but there are always open-source copycats. A lot of these models get open-sourced in a relatively quick period of time. Maybe you can't do dangerous things with Mythos because it's closed-source and Anthropic controls it, but what happens when this thing is open-source and anyone in the world can use it?

I don't know if you think about this in your personal life at all, or if you're just like, “Look, everyone's screwed. Live and let live.” I've been talking for way too long. What was your first reaction? What are you doing?

My first reaction was, “I don't have time for this.” I don't have time to go delete my search history, my bookmarks, my medical records, every photo on the cloud, every message on every platform.

There are 2 futures here, right? By the way, I think this is the bull case for Ledger. Pascal, the CEO of Ledger, came on here maybe 2 years ago and said, “We are going to secure not just crypto, but everything because nothing will be safe.”

I was kind of like, “Okay, buddy. Yeah, yeah, yeah, sure.” It was a good talking point to go do your fundraise. But I think he was spot-on here. You will need a local drive to secure everything, with the assumption that if you don't secure it locally, it will probably get out.

I do think society will change. I think it will become more normal for people in powerful positions to have more embarrassing things out in public. Think about the position of the president. The president used to be this extremely buttoned-up thing, and candidly, Trump broke that seal, for better or for worse. I'm not going to get into that here, but I think that will happen to a lot of people.

I think it will be normal to see the CEO of a Fortune 500 company partying and smoking weed or something.

We already see that.

But I think it will accelerate so quickly. It will be like, “Oh my God, crazy.” People will get canceled for a little bit, and then it will just become the norm and people will get really bored.

I tell you, your Dana does crisis management very well.

Yeah, her business is ripping right now.

There are a couple of things I think about this. One, I did go out and think about that. I saw that post. Again, I heard it many, many years ago, and I was like, “I just assume you're going to get hacked and live your life and operate in that manner.”

I think that's always a good mechanism: assume that whatever email you send, whatever message you're sending, and how you conduct yourself in private should be under the assumption that it will be public and that you're being observed and watched. That's the reality.

When you're using a social media platform, they know a lot about you. You're trusting them, and so there's no semblance of privacy today, much less going forward. We've decided that we get so much more value from being online.

Think about every call. I now just assume that every single call I'm on has Granola running in the background. Three years ago, there was no recording of calls. Two years ago, or a year and a half ago, there was a third person who joined the Zoom, and it was a bot. Today, it doesn't even join the Zoom with you. It just runs in the background.

Every call is being recorded. I use Granola for everything, and everybody does. The world is uber-connected, and you have smartphones that act like video cameras and recording devices everywhere you go.

Is the world becoming a better place and a safer place?

In some ways, yes. In some ways, privacy is an inalienable, fundamental right. But we've traded on that to hopefully get something more. What that more is and how much more, unfortunately, the cat's out of the bag. We're not going back to a world where there is more privacy. In fact, there will be increasingly less and less privacy.

I think your point is right. It's interesting because there's a Scandinavian country—I think it's Norway—where tax returns are entirely public. You know how much your neighbor is making.

It would be interesting to study that and what happened the first year they implemented it. I think it was pretty tough for people, and they probably panicked. How effective that's been is TBD. Does that create this sort of utopian society where people just—you're not supposed to talk about money, but somehow people can see your tax returns?

I've heard that a lot of Scandinavian folks have had to leave because it's a very tribal society and they don't like people standing out. A lot of the most successful business entrepreneurs— in fact, I think the Dune founder lives in Singapore—have publicly talked about certain government policies, but also about society. I've talked privately to a lot of folks from Scandinavia.

He moved to Switzerland, not Singapore, but yes.

But it's very tough to live in a society where everyone knows how much you're making. It's actually looked upon poorly if you're making a lot of money because you're not supposed to do that. It's an interesting study of what happens when everything's public.

For all intents and purposes, people will know much, much more about you going forward.

Yeah, it's true. Look at what's happening on-chain. You can quite quickly—if you know someone's address, you can see a lot of it.

2. Anthropic Announces Mythos

All right, let's wrap up by talking about inside baseball on fundraising. We talked about it a little bit last week, but do you have any more color to share on what's happening in the fundraising markets? Or I can talk a little bit about—

The nice thing about this is that we don't have funds, so there's a little bit more liberty with which we can—

Rob's like, “Every company is doing really poorly. Everything's down 80% except for Drive.”

Okay.

Okay. There's obviously a public market. I tweeted about the state of secondary markets in crypto, which I've always found to be a really good indicator.

My point was that there's an 80% to 90% discount on things that historically would have cleared at a 60% discount. It was selling the full lot. Monad, for instance, is doing pretty well. It's probably one of the best-performing, if not the best-performing, projects over the last 2 quarters.

A lot of things are trading at an 80% to 90% discount. The point that people, including Rob, are making is, “Well, it's not 90%; it's actually 60% to 70%.” But if you read through the article—and I was talking to Omar privately—this is just bids. The bid-ask spread is so wide right now that there are very few transactions happening. The bid is at an 80%-plus discount, and the ask is, if you're selling, hopefully at the historical 60% to 70% discount.

No one’s buying there. I think the biggest change in the last couple of weeks is that the ask is starting to come down. You had this spread of 70% to 90%; I think the ask has now come down to 85%, and things are starting to clear, from what I’m seeing.

Yeah, yeah.

Eighty-five percent—that’s my point. Eighty-five percent is very high.

No, no. I think it’s the bleakest fundraising market it’s been in crypto since 2015. I think 2018 and 2019 were better, and I think 2023 was way, way, way better. Way better. And I’m talking about tier 1.

There are a lot of good teams right now that have raised between $5 million and $30 million. These are kind of Series A businesses, and they’ve been selling for several months. They weren’t getting a price when they were trying to raise, then they pivoted to selling, and now they’re just selling.

I know one business today—I just talked to the founder—and they’re selling for a very small amount to a chain. That’s the other dynamic. I think a lot of chains have realized there are only 2 paths forward.

One is that we can pivot to enterprise SaaS and sell a chain in a box for $200,000 a month, try to get 20 enterprise deals, and go down that path. That’s what some of the chains are doing. I think they might do it well.

But the other path is to do what Ripple has done and start using the gift you’ve been given, which is free money. You’ve spent that money for years on developer relations, ecosystem grants, and conference parties. I think people are about to start saying, “Let’s go become a roll-up vehicle.”

That’s what I’m seeing a lot of. If you look at 2021 and 2022, the exchanges were the ones making the acquisitions. Now it’s the chains making the acquisitions. Polygon is doing that.

The problem is that it’s a very liquid currency. Why did AOL acquire a company 10 times its size?

Illiquid or liquid? It’s not a liquid instrument.

No, yeah, it is. It’s a token you can sell. But who’s going to buy the token? I mean, there’s liquidity for a lot of these tokens to trade. These are small deals.

Let’s say you raised at a $100 million valuation. Let’s say you raised your Series A at a $100 million valuation. You’re now trading at an 85% discount. Actually, it’s bigger than that: You’re trading at a 90% discount. That’s $10 million. You can get $10 million of liquidity if you just— It’s not that difficult to get $10 million if you’re an L2 token.

Yeah, over a couple of weeks, and use someone like Flowdesk and use their token.

True. Yeah. The problem is that it’s very public, and I think word gets around if you’re selling a lot of your tokens. You can’t use tokens the way Ripple is using them.

Ripple is in a privileged position, I think, because they’re smart enough to understand that they have a currency they’ve been using to acquire and absorb businesses with revenue and cash flow, like, what is it? Arabelle's and Hidden Road and a few others.

Smart. Really smart. AOL did this really well. If you have—yeah, I don’t know. I guess these deals were getting done, like OP convincing Base to come on and giving them—

But that’s not an acquisition deal. That’s just a BD deal. And, by the way, I think those are still happening in droves. Anytime you see a big financial institution come onto a chain, that financial institution did not run an RFP. They ran an RFP for a grant. They got money to come onto that chain.

I know it’s rarely talked about, but I think it’s good to put that out into the open.

Is this part of the token transparency?

Yeah, we’re working on it. We’re recording this on Wednesday, not Thursday, because I’m in D.C. all day tomorrow. We’re pushing pretty hard. There’s a lot of good stuff coming from the token transparency framework soon.

Yeah. But on inside baseball, I would monitor the secondary market. I think it’s pretty depressed. Over the last week or so, I’ve been seeing a bit more of an uptick.

I’ve also seen a bit of an uptick in private rounds. I’ve talked to a few founders and a few good companies out there. I talked to a founder today who’s in Colombia, totally bootstrapped, playing in the stablecoin space, both B2B and B2C. They’re entirely self-funded and profitable, and now they’re going out and raising a round.

The best part of this environment is that it’s instilling a survival instinct in some of the best operators. I think you said it live or backstage: There’s a discipline around optimizing for revenue, but also for sustainable growth. Focus on the unit economics.

I think that type of debate is happening all around crypto, which is a very good thing. Omega E just posted something around their TGE with pretty open disclosures. That’s what you want to see in teams: “Look, I know it’s not a perfect environment.”

That type of communication is good. This is the type of conversation that, if you remain optimistic, is worth having. It’s the right time to focus on the things that don’t get talked about in a raging bull market.

No one cares about unit economics. No one cares about price-to-fees, security, or value. Everything sort of gets put aside in a raging bull market. But this is when you can really optimize for that. It’s quiet, and if you set yourself up well—

3. The Biggest Winners in Crypto Today

Maybe an interesting, fun question for you: We do predictions at the end of the year, but I think we should start making more and more of them and then just be accountable. With AI, we can say, “Hey, how right were we over a 3- to 6-month period?”

Last cycle, what were the main winners? What were the new companies that emerged? Athena, Hyperliquid, Morpho, stablecoins, prediction markets, and perps. But some of them were already working. What were the new ones that most people missed, and then they just blew up?

Oh, new-new? New-new: Hyperliquid, Ethena, Morpho—Jupiter.

Canton? Canton? Did you say—

I don’t think that was last cycle. I think that’s new. I mean, publicly, we knew about it this cycle. I know they’ve been working on it for 10 years, but—

Yeah, yeah, I know exactly. But just the stuff that we didn’t talk about. I guess the question is: What did we talk about very little, or totally miss, that then just blew up?

And Canton was one that—I don’t think we’ve talked about it much. I had a discussion with the founder over dinner, and he gave me the full brain dump. I was like, “Wow, okay.” He was like, “Yeah, we’ve been at it for 10 years.” I was like, “Wow, yeah. I haven’t heard about you guys that much.”

I’m in one that a lot of people are talking about privately: LayerZero. But I guess that was too—

That wasn’t new. I think the new mechanism around ZRO was very closely held. We recorded a podcast while you were out.

Oh, did you end up recording with them?

Yeah, I did. They gave me insight into what was going on in January, and I was blown away. Then I brought them—

I just know of a fund that’s buying a ton of ZRO because they think it’ll just re-rate as an L1, basically.

Yeah, yeah. But, anyway, more interestingly, what are the things that we’re missing today?

We hear us talk about this, and I want to end with some food for thought for our next podcast before wrapping up with content of the week. One of the things I’ve been thinking about is how we really elevate the discussion here.

There are a couple of ideas that I have. One of them is to constantly think about what we’re missing and how we make sure that we’re finding it. Historically, we’ve brought on people who weren’t as well known and then became much, much bigger over time. Not always, but sometimes we did that with reasonable success. I want to do more of that.

Maybe we were talking about Figure. I think a lot of people are still not as well versed in Figure and Provenance as they should be, which is crazy. It’s a public company.

All right, so make the call. What are we not talking about?

I was thinking about this at the RWA Summit. I think Sky as an ecosystem—and I think you mentioned it in one of your predictions—but Sky and what’s happening with Obex is very interesting.

What is Obex?

They have these stars, and they’re trying to onboard real-world assets, but very interesting sources of yield. It ties into the discussion of vaults: How do you solve the issues with real-world assets?

We had Rune, the founder of MakerDAO, now rebranded as Sky, on the podcast, and he discussed the vision for Sky. Admittedly, I didn’t fully understand it at the time. I didn’t fully appreciate the workings and potential of Sky.

By the way, I hold zero Sky. I know other folks like Framework and ParaFi are very bullish on it, but I think it’s one of those projects that, in the conversation about real-world assets being all the rage, is probably doing the most interesting work and is the least talked about.

Yeah, yeah. I would say, if you want a prediction, RWA looping—basically, RWA leverage—will become the next vault-type thing. Is that a good or bad thing? I don't know. To your whole conversation earlier, people don't want 4%. They don't want 6%. They want to leverage to the nines.

I think RWA looping is something that reminds me of DeFi Summer a little bit. Very sophisticated people would degen into Yam and play around in DeFi Summer, but most people weren't doing that. Then it became mainstream. That's what's happening with RWA looping right now: with Superstate, there's a lot of interesting looping to be done, but it's not a mainstream product. I think it'll probably be mainstream in a year.

I remain cautiously optimistic about RWAs. Why is real estate such a massive industry? Because it has a lot of tax advantages. A lot of wealth over the last 200 years came from real estate, and the tax regime was such that, if you were selling, the tax code was very advantageous for real estate.

There's a really good podcast by a16z Perennial, which is the family office for Mark, the principals of a16z, and some of their founders. I think the sorcery podcast interviews this guy, and he talks about real estate. This is the bull case for real estate: it's very liquid. You can borrow against real estate; it's stable collateral, and then you can do really interesting things with it.

If you really want to attract a lot of capital on-chain, I think you do it with stable, uncorrelated collateral. Hopefully Michael Burry doesn't disagree with me, but it's collateral that you haven't really had before. You're still not going to solve easy in-and-out redemption. It's a legal wrapper; there are SPVs. A lot of that doesn't get solved.

One of the things that I'm quite bullish on with RWAs, as I was discussing it at the RWA Summit, is what's happening in private credit. There's this idea that you invested in a fund that pretended to give you liquidity, and then all of a sudden it lifted a gate on you. You're like, “Wait a minute.” The uber-bull case for RWAs is that you bring all of that on-chain and have the certainty of a smart contract that it will not happen. That instills a lot of confidence in a market and brings on massive amounts of capital. But then you have to figure out the smart-contract rules.

4. Are We Buying The Dip?

Godspeed. Godspeed. Godspeed. That's the takeaway of the episode: godspeed, folks. Content of the week? Wait, before we go there, were we optimistic enough? I mean, that was a pretty bleak episode. Damn.

That was pretty bad. Yeah, I know. Well, that was pretty bad. Let's keep that barometer. Maybe we can add a little barometer where an AI marks the inflection of our voice and tells us—

Are you still buying things? Buying crypto or stocks?

Yeah, I bought a couple of things yesterday. Stocks. I went long.

What did you buy?

I bought more Western Union. I rotated some of my fixed-income positions.

Wall Street Journal crypto podcaster goes activist on Wall Street. So, wait, what was the second thing? What did you buy?

Just optimizing treasuries. Just a ladder. Nothing really sexy.

I'm looking at this ETF that got pitched to me today. It's about American exceptionalism—small- and mid-cap industrial companies. Have you ever looked at this thing?

Mhm.

It's absolutely ripping, so I'll save it for another day. But the things that I'm looking at are ways of capturing volatility. I think volatility is going to continue to be a thing. How can I benefit in my portfolio from strategies like structured products, where you benefit when volatility rises?

I didn't execute on some trades because I was waiting for the VIX to go above 30. The VIX went to 35 and the war broke out. Above 27, 28, or 30 is when I want to capture some of these structured products that I think just pay you a nice coupon. The VIX never got there.

I was watching Bloomberg last night—maybe this is content of the week—almost close to midnight, waiting for the ceasefire deadline. The ceasefire got announced, but Matilda looks at me and says, “What are you doing?” I'm like, “It's like watching Netflix. It's like House of Cards.” Yeah, logged in.

Content of the week?

The greatest story is the story of the pilot who we got back.

Oh, the heart. You want to say it?

Yeah.

No, go ahead.

No, no, I've been talking too much. Go ahead.

I didn't even know about the heart.

The heart? Well, isn't there this technology that came out? They discovered where the pilot was because they recognized his heartbeat from miles and miles away.

I didn't even see that. I just know we did a crazy operation to get this guy out. I didn't know about the heartbeat.

I think the way they tracked him was with this device, because apparently the CIA has this crazy technology to track heartbeats.

That's nuts.

Yeah, that's cool. All the sci-fi stuff that was going on.

You're joining me on this DoorDash train. We're piling into DoorDash.

DoorDash? Why?

Go listen to the podcast with Tony Xu, the founder of DoorDash. This guy is the GOAT.

Interesting. I remember living in—didn't they start in Palo Alto, at Stanford?

Yeah, they did. They started as Palo Alto Delivery.

Yes. I remember being in Palo Alto when they were just getting off the ground.

They are Stanford MBAs. They were doing deliveries. It's a phenomenal episode. It's a very good left-curve, right-curve situation: you do all your analysis, and I look at what I spend money on.

This is the second time in history that I've spent more money on a service every single year for the last 5 or 6 years, and the chart is just up and to the right. The last time that happened was Amazon. For every year over the last probably 10 years, I spent more money this year than I did last year on Amazon. That's happening right now with DoorDash.

Because you're entering your dad phase and you're too lazy to go to the store?

Yeah, it plays a role. But I also think one clear outcome of the AI trade is that all the SaaS stocks got hammered. What's the opposite of that? You have to invert, Santi. What operationally heavy companies will disproportionately benefit from AI?

DoorDash—and probably Uber—are about as operationally intense as anyone else. I'll save it and probably write a post about it and talk about it here, but I have a couple of things I'm looking at: capturing volatility. We're going to be in a persistently high-volatility environment.

People talk about this broad basket of AI-resistant companies, which includes some of the industrial companies you were just talking about. There's also reshoring, so U.S. industrial companies probably bode well. I think a lot of them have just run up quite a bit.

The other one is real-world experiences. If the world is only going to continue to be more stressful, you're going to pay a premium for disconnecting. The watch industry, when the Apple Watch came out, only ballooned. You know how much I love watches. You're going to crave the nostalgia of analog—of disconnecting from all this noise, craziness, and stress. Your cortisol is high, so you want real-world experiences.

Maybe it's amusement parks. Maybe it's luxury travel. People will crave disconnecting more and more—forcefully disconnecting—because social media will suck you in. As it relates to my kids, there's no social media. It's very, very destructive.

To wrap, I'll write about it and we'll talk about it. I have a content recommendation. Everyone should drop everything and go listen to these 2 podcasts: Invest Like the Best and The Knowledge Project. Both bring on Jim from Alpha School.

Oh, so good. So good. I signed up for their—

Did you get in? They have one in New York?

They have one in New York. I'm on the list. I realize they don't take tiny, tiny babies, but I'm on the list just to see how it develops, and I get their newsletter and stuff like that. I'm dying to get in.

Schools in New York are so messed up, by the way. You have to literally start networking when your kid is 6 months old to get into daycare, preschool, and kindergarten, unless you're zoned for a good public school.

Yeah, Alpha School is very exciting to me. I think it's a great concept. I've never felt—

I DM'd him on Twitter. I was like, “I need to bring Alpha School to Monaco.” There just isn't enough school here, either. If anyone in the audience knows him, get me in touch.

My first business was tutoring. I think education has been broken. I've written about it. The simple solution isn't what Mark Andreessen goes out and says, where a kid that has been tutored has two standard deviations of performance on standardized tests like MAPs.

The lazy assumption would be, “Oh, you’re going to prompt, and your kid’s going to be a savant just because it’s so engaging.” But he actually came out with a little bit more nuanced answer, which is: if you think AI is going to be the solution to education, you’re very, very wrong, and they’ve been at it for 10-plus years, I think. It’s just remarkable. I’m like, where have I been all these years? Why have I never heard of Alpha School? It started in Austin.

Yeah, I agree. That was a good one.

Are we going to plug DAS London? Dare I say—how is it? October?

We're plugging DAS Abu Dhabi. They're booking the venue for $2. They're paying you.

Hell of a year to push into the Middle East, my man. You thought vaults had risk, man. To be fair, it's an industry that craves risk. Let's go, CertiK researcher number three. We're plugging DAS Abu Dhabi. Not yet. DAS London. DAS London is live. If you want a good sponsorship booth and want to get in early, we're selling meetings with Santiago for 25k. It's a bear market, folks. Bear market. You know it's serious when you use my full name, man. Santiago Roel Santos. October 20th and 21st. Be there.

5. Content of The Week

Content of the week is this Tony Xu episode.

I think it’s a tough time for founders right now in crypto, and for me personally, it’s inspiring to listen to people who’ve just been through the absolute trenches. Oftentimes, you think you’ve chewed glass or been through it, and then you listen to another founder who’s been through it worse. He has amazing stories: the big fundraise is about to close, he goes on a honeymoon, the market changes, investors pull, and he can’t raise money for 3 years for DoorDash, a very financially intensive business.

So you remember I told you the reason—probably the single biggest reason—I started Version was because I listened to 100 hours of founders’ podcasts, and I thought, every major founder goes through a near life-or-death experience with his company. It’s never easy.

I’ll get you connected to David.

Oh, man, 100%. I’ll be his patron. He’s a legend, man. I now see that he’s going on other podcasts, which is great.

The man’s as intense as it gets. I love it.

Yeah, listen to that. Honestly, that is the closing message. The book I was reading was Barry Diller’s “Who Knew?”

Yeah, I read that.

He talks about a really interesting concept. He was like, “I overcame business fear because I had a far greater personal fear about my personal life that I had to hide for a long time.” You can imagine what that was. That made me think: could you manufacture a greater risk to overcome the risk of whatever it is—starting a business? I think that kept him delusional and optimistic for a real long time.

And I think, yeah, it’s tough, but then you break a finger and you’re like, “Holy shit, I would give everything for my finger to be good.”

Yeah. So if you’re healthy and able, it’s not that bad.

Yeah, exactly.