市场更新：Situational Awareness 如何爆雷，利率会把我们炸毁吗，接下来是什么？

The lesson here is that the worst thing you can do in trading, or frankly in life, is to be dead right and sitting on your ass on the sidelines with no exposure because you got kicked out. You got marginalized somehow.

1. Did Ken Griffin Kill Leopold Aschenbrenner?

Crazy day in the market. Amazon up 15%. Google up 6%, Apple down 10%, Leopold blowing up, the Fed hawkishly not doing anything. This is a hell of a market. This is the greatest game ever. This is finance at its best. Did Ken Griffin kill Leopold Aschenbrenner? Are we done with this sell-off? Are we going higher?

You know, I'm really glad that it didn't happen by accident on Wednesday. I hope you're okay, by the way, because a lot happened yesterday. What a wild, wild market we are in. First of all, we have to talk about Leopold Aschenbrenner. No, Ken Griffin didn't kill him. He killed himself.

Ken Griffin didn't help.

No. I dispute that. I say Ken Griffin helped him more than anybody else helped him. I heard Millennium bid on his portfolio as well. Citadel paid the best price, so they helped him out, frankly. He would be worse off if he hadn't had Ken Griffin in the world. Ken Griffin paid him the best price.

But this is a story as old as running 10x leverage on FTX and getting liquidated, or BitMEX. People were treating Leopold Aschenbrenner like he was Jesus Christ because he was a nerdy guy who had a good academic pedigree and a sort of pedigree on LinkedIn working for OpenAI. It's like, okay, AI is the thing. This guy uses AI to trade.

Let me tell you who uses AI to trade effectively: Ken Griffin uses AI to trade. Renaissance Technologies uses AI to trade. Two Sigma, Jane Street, and HRT—they use AI to trade. You don't need to be a 22-year-old whiz kid with a weird laugh to use AI to trade. There are pros out there who are a little less well known by the people on Twitter, a little less memeable, but they're better traders with AI.

I think it's kind of ridiculous that this guy literally just used his reputation to raise billions of dollars and YOLO'd it all into memory stocks, which still would have been fine if he hadn't leveraged himself to the hilt and gotten impaled. And of course, he got stopped out on the lows, because that's always what happens.

I read somewhere this morning that Ken Griffin is probably already up $3 billion or $4 billion on the trade. No, Ken Griffin didn't liquidate what he bought from Leo. Ken probably knew he was liquidating this guy on the lows, and his marks are probably way lower than where it was trading on screen when he liquidated Leo. He had plenty of cushion to hang on, and it would have been ridiculous of him not to. I bet those books are up billions today—$3 billion to $5 billion at least.

2. How An 11-Figure Liquidation Actually Works

I mean, look, the way that this works is a lot more complicated and a lot simpler than most people expect. Did Ken Griffin himself—first of all, did Ken Griffin himself do anything? Probably not. Ken's probably on a yacht somewhere sipping champagne. I don't know how active he is in these types of situations. He used to be extremely active. I don't know how active he is.

These guys get involved in the big liquidations. At TRW, when we were liquidating, we were bidding on—

DRW's sort of a different beast. Do you think Ken is genuinely involved?

On something of this size, the CEO of the firm will get involved and sign off on it. But no, he won't be deeply involved in the pricing. He'll trust his guys. He'll just take a final scan and see if there's a problem.

3. Why Avi Is Buying The Memory Dip Now

Look, there are all sorts of questions here. But the key point that I think you were going to make is leverage. It's 3 things that will always sink you, like Charlie Munger said: liquor, ladies, and leverage. Avoid these 3 things and you will otherwise be a happier and more verdant human.

Genuinely, these are extremely tempting things to take advantage of, right? You look at somebody like Leopold. Let's not mince our words: the man is a genius. He wrote a paper 2 years ago called “Situational Awareness” that outlined effectively everything that's happening right now in the markets. He mapped out how big AI was going to get before the vast majority of people saw it.

But the problem is that winning begets arrogance. I've experienced this, too. This happened to me at BlockTower, where we ran up $40 million to $1 billion. I felt like a god. I remember, for the first time, watching Bitcoin rip from $25K to $50K, and we were—I had levered the firm, not actually levered, but we were, I think, 130% net long if you include the delta of the options that had ended up in the money.

I'm thinking to myself, “I'm invincible. I'm brilliant. I'm amazing.” And that was so much smaller than what Leopold had accomplished. Sure, we made a couple hundred million bucks, but Leopold made billions and billions of dollars on a unidirectional bet basically saying AI is going to eat the world. And guess what? It was, and it still is.

But what he did was get a little too comfortable with winning. And he didn't have senior risk managers around him telling him, “Hey, if you're going to long assets that are 120-vol and you're even 2x levered, you're going to get into trouble. It's going to happen.”

This is a lesson basically to everybody out there to be careful with the amount of leverage that you put on, whether it's in crypto, whether it's in equities, whether it's anywhere. You can be right, but leverage makes your bets entirely path-dependent. If you're going to take long-term bets on the future of the world, you kind of need your bets to be path-independent, especially when they're high-vol. You can't put yourself in a situation where you get taken out before your theory actually materializes into reality. That's really the biggest issue.

Now, I think Leo, again, obviously, is incredibly talented. He's going to bounce back from this. The fund, contrary to popular belief, is not done. It's still here, and it's still going. Actually, the TBPN guys got a letter that he sent to his LPs. I want to read an important part here.

“The fund was not shut down, liquidated, or transformed into a private-only fund. We continue to operate as a hybrid public-private fund as before. However, we will manage our public book on a fully paid-for basis while we draw the lessons from these developments. Most importantly, we took the steps that were necessary to fight another day.”

4. Amateur Hour: Why He Broke The Cardinal Rules

That last line is the key: “We took the steps that were necessary to fight another day.” That is what you need. That's the mentality that you need in order to be successful in these markets. You need to—

I agree.

Taking a broader step back, vol is through the roof. Amazon—Amazon, Jonah. A multitrillion-dollar company—is up 15% today. 15%. Microsoft went up 15% in a day just, I think, 2 or 3 days ago. What's happening?

What's happening is that the markets are becoming far more volatile. This is one of the reasons why crypto can't really catch a bid, despite what I viewed as strength and now I view as almost a missed opportunity, which we'll get into. I think what we need to start covering is how to deal with these high-vol assets now. It's totally insane.

Anyway, I'll stop my rant there because I want to hear what you have to say. I do want to dive into whether Ken Griffin orchestrated this, or whether this was just a natural extension of what he was doing.

No. I mean, look, I think you touched on something very important in what you just said, which is that if you have a long-term thesis, you probably shouldn't express it with 120-vol stocks by YOLOing $20 billion or $30 billion into those stocks with 4x leverage.

To me, the reason why I watch this—and I've tracked Leo a little bit; I've watched his lore develop on Twitter. I haven't said anything because I don't know whether he is who people made him up to be or not—but to me, I'd look at this and this is just amateur hour, right?

4x leverage isn't the end of the world if you're a retail trader in crypto or just trading levered stocks on E-Trade or Schwab or whatever's out there, Fidelity.

But you lose your ability to dance between the raindrops when you have a 12-figure position, right? When your position is worth more than $100 billion, you can't get out and get back in, get out—you’re too big. You're the market, right? So this should be obvious. This is amateur hour not to do this even with small dollars, but with that amount of money, you definitely shouldn't be doing this.

To me, it seemed a little weird, this guy raising so much money. It seemed like he was raising the money and, on vibes, trading it based on something he put on Substack, which, let's give him his flowers, was probably prescient. But if you make a long-term prediction, you nailed it, obviously. You cannot trade liquid, short-term instruments and just set it and forget it with 4× leverage on $30 billion of capital and $120 billion of market exposure and just hope for the best.

You need to not be stopped out before your long-term thesis plays out. It needs to be a bottom-drawer trade where you don't have to care. To me, this is, again, I'm not going to grave-dance because he's probably still got liquid positions in Anthropic and 11 Labs and whatever else that he's run up, and he'll probably still be fine. Just like people were grave-dancing my former colleague Yao Yao when he had his speed bump earlier this year, he's going to be fine, too.

All these guys are going to be fine. But unlike Yao Yao, who has crushed and made billions of dollars for decades now, this guy shows up and then a few months later gets liquidated. No, Ken Griffin didn't liquidate him or orchestrate this. Ken Griffin just showed the best bid to pick up the pieces after this guy made his own mess in a very, very immature, basic way.

So I guess the lesson for the rest of us is: for a long-term thesis, 2× leverage at most. Make sure you're going to be okay even if the market is irrational. You have to stay solvent somehow. And the other lesson is that there's no back room or cabal or Eyes Wide Shut party where Ken Griffin and Izzy Englander decide that Leo's going to die, right?

The dude just blew up on leverage because we were in a white-hot bull market. I'm personally buying a little bit of Micron stock here. I'm dabbling. I wanted to see a bottom and a bounce, and I think the market picture is much cleaner now. I think that, realistically, Izzy and Ken and anybody else who bid on Leo's book recognize that this is a bull-market positioning unwind.

It is not a sell-off because the AI thesis has been disproven. It's pure technicals.

That part I 100% agree with, and I want to get there, but I want to take us on a little bit of a journey first. When you talk about Leopold and how he constructed his book, the reason that we're both sitting here and saying he constructed his book in a reasonably amateurish way is because a lot of people in the hedge fund world go to great lengths to conceal their positions. If people know what your positions are and they know that you're leveraged, they will hunt you.

It's very easy, when you have somebody like Leopold, for pod shops to figure out what positions they hold because he's publishing them publicly, and he wasn't doing anything particularly intelligent to obfuscate his positions. Sometimes, what Long-Term Capital Management would do, as an example of this—because it's public and it was in a book—is that they would actually take offsetting positions with different brokers, some slightly larger, some smaller, so the brokers' sell side talks a lot.

One thing the sell side can't do is give you exactly the amount that you're long of something. Somebody who works at Goldman is not going to talk to one of their clients and be like, “Hey, by the way, that other firm, your competitor firm, they're long $1.2 billion of this stock.” All they're going to say is, “Yeah, those guys are bullish on that stock.” There's sort of a level to the amount of information that they can disclose, and people obviously talk—actually more than you might think.

For example, you might take offsetting positions with different brokers. So if you call up Credit Suisse, they might say, “Hey, Aschenbrenner's long this stock.” And you call up Goldman, and he's like, “Actually, he's not—no, he's not really that long that stock. He's actually kind of short it.” And so you try to obfuscate some of your positions so that they don't come out on the street so obviously, and people have to really pay attention to your filings and whatnot, which only come out once a quarter. But they can't get updated positions.

That was a big mistake on his part: he didn't spend enough time obfuscating his positions. Of course, we're just kind of nitpicking here. The dude's made a ton of money. He's still up year-to-date. He's still up 80% year-to-date, which tells you just how well he was doing despite having a 67% down month and getting margin-called, still up 80%. That's impressive, but I'm just walking you guys through the mechanics.

Do you really believe he's still up 80%?

Maybe he was up 80%.

But when did that come out?

That's what TBPN published today.

But the investor letter was from a week or 2 ago.

No, the investor letter was from today.

Okay. I don't buy that.

I buy it. He can't really lie in that unless it was a fake letter. Obviously, it hasn't been corroborated yet. Maybe I'll wait for it to appear in the Financial Times, because that's real journalism as opposed to this fake journalism, which, obviously, I classify as fake journalism as well. So, no shade.

You can be a fake journalist. You can just say stuff.

It's so much better to be a fake journalist. Once you go behind the scenes and figure out what journalists actually have to do in order to put something in print, it's just so much work. At the end of the day, some of it still ends up being totally wrong and dumb.

I'm thinking about The New York Times and how they actually make people jump through hoops to cite sources for things that end up being completely incorrect. They make extra work for no reason sometimes. So I'm just glad that we're actually not journalists.

5. Walking Into The Lion's Den (And The Luna Parallel)

Anyway, that's a totally different rant for a totally different time. I'm getting way too sidetracked. The main point here is that the way Aschenbrenner blew up is really, really, really simple.

He took on far too much leverage, and when the trend started going against him, when there started to be cracks in the market because people started taking profit, the entire street started smelling blood. Then he made a fatal mistake. He went to the street and tried to raise money to meet his margin calls.

That is like going into a lion's den with a piece of meat and saying, “Hey, does anyone have any extra meat to spare? I have a lot here, but obviously the lions are going to attack you.” Obviously, they're going to attack you. They're not going to just take your meat. They're literally going to hold you at gunpoint and take it from you. What are you doing?

You have to be smarter than that. You cannot go and broadcast your weakness to a bunch of sharks or lions or tigers, or however you want to characterize these Wall Street investors. Obviously, they're going to short these stocks, and you're going to get even more pressure on your positions. And that's kind of what happened.

This is the same fucking trade, by the way. The same trade as why Bitcoin goes down in front of Saylor. Why Bitcoin went down when people saw that Saylor was publicly stressed and stretched, right? It's the same trade.

The only difference, Jonah, is that it happened behind closed doors. You, the retail investor, you, the person who isn't on Wall Street, cannot and will never know that this stuff is happening until it's too late. You have to be good at reading the tea leaves and figuring out what's happening in the markets.

When you get moves like what you saw in memory stocks, where things are going down 15% in 1 day, you have to start thinking to yourself, “Is somebody getting carted out?”

Yeah, that's the key. Right? You know what happened? This reminds me exactly of my call with Do Kwon in 2021—or 2022, sorry, like summer of 2022—when he went behind closed doors and called me at Cumberland and a few other of our cohorts. He was on the Zoom with Kanav Koria, the mustache warrior, and they were trying to raise money to shore up basically their Leopold Aschenbrenner hedge fund of crypto, where they were just taking funds and yoloing it into Bitcoin.

Remember the LFG foundation of Terra Luna? It would just buy Bitcoin with Anchor Protocol money. Anyway, basically, he called and he was like, “Hey, I need $100 million. Can you get me that and I'll—”

You need $100 million?

No, we were like, “How much do you need in total?” He was like, “Well, we're going to raise $1.5 billion. We've already raised like $800 million. Do you want to chip in $100 million?” And we're like, “Whoa, how'd you come up with $1.5 billion, and what are you going to do with it?” He's like, “We're going to buy Luna.” And we're like, “Well, how'd you come up with $1.5 billion? Is that the right number to spend buying your own token?” He was like, “Well, we had just—you know, he had a bit of a lisp.

He was like, “Well, we looked at all the offers on all the exchanges, and that’s less than $1.5 billion. So, if we have $1.5 billion, we’re going to take out all the offers and scare all the sellers, and then the token’s going to go up and you get your money back.” And then some.

I was like, “Hmm, what if there are other offers that aren’t literally just on the exchange? Like an iceberg order? You don’t always—when you want to sell something and you put in a limit order, you don’t always show your full size. Have you thought about that?” He’s like, “Yeah, we thought about that, and it seems unlikely.”

I just remember the answer was obviously, “Go fuck yourself.” He walked into the lion’s den with a piece of meat asking for more meat. We went and ate the meat. We shorted Luna. First, we sold what Luna was selling.

That is why this is the greatest game on the planet.

Yeah. You have to think adversarially if you’re going to win. There’s no world in which you can approach this as a bright-eyed, bushy-tailed guy and say, “People are on my side and we can all make money together.” Somebody is out there, and they’re trying to fucking kill you if you get big enough.

If you’re small, don’t worry. You’re worth $1 million, you’re worth $5 million, you’re even worth $100 million—nobody’s going to care. But if you’re out there running $25 billion and you lever up, somebody’s going to try to fucking kill you.

I bet you the way Ken made his money on this—I bet you Ken shorted Micron. I bet you Ken saw this, took the call, shorted Micron stock, and the reason why he was able to be the best bid for the portfolio is because he was short. I bet you that’s what happened.

Absolutely. I think there are 2 reasons why Citadel was the best bid here. One is that I think the bidding was Jane Street, Millennium, and Citadel. Jane Street’s not going to be the best bid, kind of ever, for this kind of stuff. Why? Because AI has a long-term time horizon.

If you have any inclination to hold this position for more than 38 seconds, Jane Street’s not going to be the best bid, because Jane Street is just thinking, “How can I buy and liquidate?” They’re a trading shop. These either have to be complicated-to-price assets, or they have to be complicated to price. Maybe they’re debt instruments, or some weird convertible arb might be going on there.

If they’re complicated to price, if they’re a bunch of structured products, then a Jane Street or Hudson River Trading—these trading firms—are going to be really good at that and be able to price it really well. But these are just straight long equities that have a very simple thesis and somehow were held by this guy who needed to meet margin calls.

It’s not as complicated to price Bloom Energy. It’s not as complicated to price outright equity of Micron or outright equity of SanDisk. It’s not necessarily something that you need high-IQ, high-functioning quants who can’t speak English for. You actually can have English speakers do that. And so what I’m trying to say is that Jane Street is probably not the best bid for this type of auction.

Right. That’s what we’ve been saying, too.

And he’s correct. He’s correct. But that makes him naturally a better bid for these long-term assets, or assets that could theoretically be construed to have long-term theses tied to them, than Millennium, where Izzy Englander is still more focused on short-term trading.

I actually think that the risk limits at Millennium are tighter than the risk limits at Citadel in many pods.

Way tighter.

Way tighter. And so that’s why, on this type of fire sale, Millennium or Citadel is going to come in. If you had told me a priori, “Hey, these are the 3 people bidding,” I think I could have guessed Citadel with 90% accuracy. Obviously, I didn’t know, but there’s only one firm—and this is totally irrespective of Ken Griffin getting on and calling for rate hikes.

Do I think that this was an orchestrated hit? Because that’s the question that we keep getting. When we’re done talking about this, we’ll talk about why it matters. It absolutely was an orchestrated hit, in my personal opinion. People smelled blood in the water and attacked. They did whatever they needed to do to attack.

I mean, did Ken Griffin go on national television and say that he thinks rate hikes are coming with the explicit belief that he would be able to buy Archegos’s portfolio? Probably not. But did he have it in his mind when he was doing that that it might be helpful? Probably. That’s my personal opinion.

It was probably in his mind somewhere that this was a stressed market and that by pushing it, he might be able to break it, and they might be able to make billions of dollars. Now, I don’t think he would have gone on TV and burned his credibility saying that unless he somewhat believed it. But we’re not in his mind.

Let’s say you’re a manager and you believe that in the next 12 months, rate hikes are going to happen. He didn’t frame it as rate hikes are happening in 2 weeks. He just said rate hikes could happen. Let’s be nervous; the market maybe should be nervous about rate hikes.

But why say it now? Why, at that moment, go on and talk about it? Yes, it was before the Fed, but I think it’s because you wanted to put some pressure on the market. That’s my personal opinion. I have no idea.

Now, obviously, this leads to a mechanical blowup of the Situational Awareness Fund. So, let’s talk about what to do. Those types of forced liquidations are often the end—the end, right?—of the total collapse. But they’re not necessarily the end of the selling.

Why? Because after that, you get this mechanical bounce. Things are up 25% or 30% off the bottom in some cases. Today, you actually saw SanDisk up almost 40% from the lows. Then it gets smacked down.

Why does it get smacked down? Because, A, there are people who bought the lows. B, there are people who are massively short and are covering their shorts. And then, obviously, Citadel has to take some risk off the table.

Yeah.

That’s why, in my tweet yesterday, I said, “I think that we’re probably in for some churn.” Then the real question is, “Is it possible to get back to the highs? Will we trade higher, and how quickly?”

Now, the answer you can find by diving into the question. If we only got to the highs because of LL, right? Leopold. If we only got up there because of Leopold leverage—that’s what LL stands for; it’s a new term I coined, the LL—then we only got to the highs because Leopold added $25 billion to these companies.

He borrowed $25 billion from the banks and shoved it in. So that’s how we got up there. That’s why July was such a crazy month for these assets. It’s because he was shoving it in on these names. Now, if he’s wiped out, he’s probably not doing that again.

So is there $25 billion of capital outside that’s willing to come back in and basically pick up the pieces of what Leopold left? My answer to that is: it’s very likely yes, but not yet. It’s going to take maybe some time—maybe 2 weeks, 3 weeks, 5 weeks. I don’t know exactly how long it’s going to take.

Capital’s going to be shifted around right now. You can see it: capital’s in Amazon, up 15%; capital’s in Microsoft, up 20%.

Yeah.

I think it’s going to take some time for the memory stocks to go back up, but the reality, Jonah—the reality is that we’re in a secular megatrend. And so I think this is actually a pretty reasonable time to get in.

This is what I was waiting for. I’m in.

Yeah, this is the moment, I think. I tweeted this out—I think this was 10 days ago at this point—and said, “Look, this is maybe a good time to start nibbling on memory stocks if you haven’t been in.” Then I said, “I would be all in 20% lower.”

When I said that, DRAM was trading at about 57, and 20% lower from 57 is, let’s call it, 47 or 48.

Yeah.

We bought, went to 44, and went lower. But I think we’re still in that area. I think DRAM’s trading at 51 now. This feels like the right time to start reallocating to memory, especially Intel, which was one of Leopold’s largest positions and still obviously has those massive national security implications for the U.S.

I’m now more allocated to memory than I’ve been since I told you guys—I think it was 2 months ago at this point—that I was starting to get nervous about the memory market and had moved away from the memory trade to the biotech trade, to the other downstream trades.

Now, candidly, guys, the biotech trade hasn’t done incredibly well. I think ARKG’s up maybe 6% since I moved into it. XBI’s actually down from when I bought it.

So, that was an L. BLLN (Billion to One) is up actually quite a bit. I think it's up 20% since I mentioned I was moving over, but my biggest position was XBI, so I think net I'm up, blended, 3% on these positions, but at least I'm not down massively. My Intel—I didn't fully exit, so obviously I went down on that, but now I'm sort of reallocating my cash back into memory here.

I did a little bit yesterday. I'm going to do more today, but I do think that you can accumulate a little bit more slowly because it's possible we test the lows again before we go higher. Now, the names that I'm buying: just RAM, Intel, Micron, SanDisk. That's it. I'm not touching anything else.

I'm not touching SK Hynix. I don't know the Korean market. In fact, they missed earnings. That was going to be tough, I think. That was going to be a tell for the collapse of the market, and it's kind of tough because the market's already collapsed. The majority got wiped out, and then they missed earnings, so I would stay away from South Korean equities.

I read something crazy. I read something like 5% to 10% of South Korean adults—of the entire adult population of South Korea—got liquidated on this. Something hilariously wild, straight out of Squid Game. It's like one of those games where a bunch of Korean dudes are trying to run across a little plank over a giant abyss, and 5% fall off. That's basically the Korean economy just got Squid Gamed.

Back to what you were saying earlier: the reason why I'm getting involved now in Micron, and I'm intentional about Micron, is it's here in the U.S. of A. I like it. I do think memory as a physical commodity is undersupplied by at least 40%, probably 50%, for the next 12 months. I think that the price of those modules—the underlying physical—is going to triple in that time. That's good for Micron. I'm not sure about Hynix and Samsung; you can keep it, but Micron's safe, especially at these levels.

You're buying a dip in white-hot bull-market volatility. You're not buying a dip in crypto in 2022, when it seems like the technology may go away. This technology is not going away. So I do think this is a good time.

Back to what you were saying about Leo, and then I'll finally shut my trap on that topic. I think that Ken Griffin threw stones, but I don't think he is the cause of Leo's downfall. I really—I'm going to say it again—Leo caused Leo's downfall on his trade.

When I worked at Goldman, when I worked at Vitol, even when I worked at Cumberland, people were always like, when there were big trades going through in the options market in 2011, everybody in the options pit—which still existed back then—would be like, “Oh, it's Goldman doing the trade.” Everybody wants a boogeyman. Everybody wants a guy or a team or a firm to assign the big stuff to.

In the case of Bitcoin, Saylor may actually be that boogeyman, which is why I hate my Bitcoin holdings so much right now, even though I'm not liquidating them because I have a long-term thesis. But I don't think Ken or Leo was 100% of this AI trade, right? So basically, I think Ken threw stones. I think Leo ran it up a little bit, but to the point you made, I think it goes straight back to the highs without them.

6. The Real Lesson: Being Dead Right And Broke

I do not think that Bitcoin can go straight back to the highs right now without Saylor getting liquidated. So it's very different; it's apples and oranges.

Now, to the point about what I think the biggest lesson from this is: it isn't leverage. It isn't amateur-hour-style portfolio construction. It isn't, “Hey, if I have a long-term thesis, maybe I shouldn't put on a short-term trade.”

The lesson here is the worst thing you can do in trading, or frankly in life, is to be dead right and sitting on your ass on the sidelines with no exposure, right? Because you got kicked out. You got marginalized somehow. Leopold was dead right about this trade, and I am personally stepping into his poorly expressed trade without leverage myself because of how right I think he was and is. But I'm not going to get booted out because I'm not going to put on leverage and get liquidated by my prime broker or my trading platform.

Being right and losing money is the worst feeling in trading. Being wrong and losing money, okay. If you're never wrong, something's weird. You're either not taking enough risk or you're running a Bernie Madoff-style Ponzi scheme. You're going to lose money on being wrong all the time. Losing money on being right is the worst. It's so unnecessary, and it goes back to Chris Rocos's famous truism: nothing matters more than trade expression.

It's true, and also position sizing. This is something that—

That's part of the expression, I'd say.

Yeah, no, for sure. It's a subcategory of expression, but I think the problem also is that the assets that Leopold was trading were not, in the grand scheme of things, big enough for his position sizing. I think he was 25% of NBIS. He was like 5% of some of these others, but in terms of market cap, really, you need to be looking at ADV. You need to be looking at average daily volume. Generally, the guideline, I think—

Average daily volume—

—is, yeah, it's 10% of ADV. You don't want to be more than 10% of ADV. That's the general guideline. Some of these general guidelines actually exist for a reason.

I think a lot of the time, especially in this industry, people think to themselves, “Conventional wisdom is dumb, and in order to be successful, I need to move away from conventional wisdom.” I'm here to tell you that's kind of true, but mostly not true. When it comes to statistics, when it comes to math, it's not true.

The conventional wisdom of the wunderkind almost always blows up. Almost everybody who is the best-performing fund in any given year is going to have a blowup moment. Why? Because in order to be the best-performing fund, you probably need to be taking excess risk. You probably need to be ignoring portfolio construction. You probably need to have a ridiculously high-volatility portfolio. It's very rare that an individual fund is the best-performing fund of that year and isn't taking on excess risk.

How many down years have you had in your career? I've had 3 as a professional money manager.

As a professional money manager, I've had 2. But you were managing money longer than I was.

Yeah. What I will say on that very quickly is I think the only guy who's ever done it at scale without a down year is Druckenmiller, right? And that's why he's the GOAT. There just—there's never been better, there never will be better. I don't really count Jim Simons and Renaissance Technologies because that's like trillions of little agents trading together and cumulatively making money, but in terms of just a guy with an opinion, Druck is the only guy, right? He's just the legend.

Now, I think every other amazing trader I've ever met has always had 1 epically bad year and at least a couple of other marginally bad years, and maybe this is just that year for Leo. So be careful grave-dancing on this guy. Hopefully he has a good year.

Fair, he's still up this year. He's still up. It's just the drawdown that was big. Now, I want to talk about that.

A successful venture capitalist does not make you a successful trader. We need to differentiate. So I'm skeptical of that up 80%, but if you got lucky on some locked-up Anthropic and got liquidated on your public-markets portfolio, let's separate the venture capitalist—which he may be the greatest of—from the trader. Let's call a spade a spade here: he's not up on liquid markets this year. No way.

I don't know. The answer to that is I don't know. All I'll say is that he has been—and this is the last point because I want to stop talking about Leo. I think that topic has been beaten to death now at this point. But I think that you can't count him out, and I'll be watching his career with great interest. I hope he does well.

This is not a grave-dancing podcast. We are positive supporters. We are positive people. If you have lost money in the last month, just know that you're in great company and don't feel too bad about it. There are a ton of people that have lost money.

In the future, maybe you can listen to this pod and we'll help you hedge a little bit, I think. That was one thing that we did well: we rang a little bit of the bell. I remember that story that I told you a while ago where I went to an event and there was a kid there who was buying memory stocks, and I said, “This is kind of scary.” That was actually pretty close to the top.

Now, what I need to get better at is my public Twitter and reflecting what I'm saying on the podcast, because sometimes I'll tweet something and forget to follow it up. I did tweet about Micron earnings, and people took that to mean that I was still ridiculously bullish on memory in general, but on the pod, obviously, we're talking about, “Hey, maybe we need to be a little bit more careful.”

7. Atoms Over Bits: The Mag 7 Split & What's Next

Now, what do I think is going to happen here? I do genuinely think that because it might take some time to get this rally started again in memory—because you might see some profit-taking—you can probably stay in the MAGs. I think the Mag 7 actually might do so well ex-Apple. Apple is down 10% today because they're basically forecasting lower revenue guidance for September and Q3.

Down 9%. I mean, these moves are massive. And I think what the market is starting to really come around to is that there was a moment there where people were saying, “Wow, maybe these companies like Amazon, Microsoft, and Meta are spending far too much money on CapEx, and it actually just doesn’t seem to be true in any meaningful way.” It’s like Apple is still struggling, right, despite spending zero money on CapEx. And so, I think the issue—

That logic doesn’t follow, Avi. It doesn’t follow.

Tell me why.

Apple struggling because they spent nothing does not mean that Google and Meta spending a bunch of money is not also bad.

I’m saying that, well, potentially that’s true, but I’m saying that if we generally have considered these stocks reasonably correlated, which we have in the past. Generally, Apple tends to move with Google, tends to move with Amazon, tended to move with Microsoft; they tended to have high correlations. Now, there’s actually quite a bit of dispersion in these stocks and actually dispersion from each other.

I think the point here is that Apple is doing poorly relative to the rest of the Mag Seven right now, despite spending no money on CapEx. Amazon did very well today despite spending a lot of money on CapEx. The free cash flow of these companies is coming down, whereas free cash flow for Apple is still quite high. Basically, what I’m saying is we have to unpack it: There might be something else going on. What was driving the underperformance of these stocks is the question. That’s the question that I’m trying to answer: Why was Google underperforming, why was Amazon underperforming, and why were all these other things underperforming for so long? Was it due to CapEx, or was it due to something else? If Apple’s also underperforming, maybe it’s not solely due to CapEx. Maybe there’s something else going on. And my answer is that it was a flows-based issue.

I think it’s apples and oranges.

Okay.

Pun intended. I think Apple is underperforming because they haven’t innovated anything since the iPhone, basically. The iPad, whatever. And they don’t even seem to be trying. They seem to have just admitted defeat, planted the white flag, and are just sitting and watching idly by while Siri is a piece of shit and the AI thing takes off.

They have this incredible opportunity in front of them, and they just seem kind of lost. They don’t know what to do with that. They’re like, “Well, we’re going to optimize our supply chain and cut costs.” That’s what Tim Cook molded this company to do. But they’re really missing a Steve Jobs right now, and this new guy probably isn’t it.

So, I think if Apple’s underperforming, it’s because they’re a company without a vision in an era that they have ample opportunity to participate in, and they just don’t even seem to be trying, which is pathetic and weird.

As far as Google, Google printed its first quarter of negative free cash flow in decades. And that’s why Google stock shanked. That’s a very different thing. It’s not like Google’s sitting on the sidelines, trying to not give a shit, right? They’re competing in every arena where AI matters, and they’re competing really effectively, too. Gemini’s a great product.

I think the reason why Google’s underperforming is because participants are just saying, “Holy moly, we assigned a software-margin multiple to this, basically the best business in the history of the world: search.” The margins are supposed to be enormous. What’s going on here?

Now this looks less like a software company and more like a manufacturing company, where they’re just buying tons of hardware to manufacture tokens, and those tokens aren’t necessarily bringing value back to the bottom line. Otherwise, the free cash flow would still be positive. So, I think that’s an equity-research-type situation that’s driving the underperformance of Google stock. It’s just analysts re-rating and re-underwriting the situation there.

As far as Meta, I’ll—I'm out of Meta now. I had a bull thesis on them. I made less than I should have on it and got out. But the thing about Meta is they also have an incredible opportunity in AI. I would place them more in the Apple category than in the Google category. They’re kind of in between. They’re in the middle of that spectrum, and they’re sort of trying to compete.

It should be a layup for them because they have billions and billions of people using their products every day. So, they have the easiest sales funnel to build consumer AI, and they can’t figure it out no matter how much money they spend. They just can’t figure it out.

So, they’re between Apple spending no money and Google spending infinity money. They’re spending a lot of money, but they are neither figuring it out nor are they not trying. It’s a weird purgatory, a limbo, for them. So, I think it’s a very, very bad situation for those 3 companies.

Amazon, on the other hand—happy days over there. Whatever they’re doing, they’re doing it right, and I’m not exactly sure why, to be honest.

What do you mean you’re not exactly sure why? They just keep printing money hand over fist.

Yeah, and I don’t know how. I guess what we’re learning here is that atoms are starting to matter more than bits, right?

Well, I mean, that’s the entire thesis behind Calyx's new company, right? That’s—actually, Amazon has always thought this in many ways. Even though they’re obviously an internet-first company, what they really built is physical hardware in AWS, and they built logistics to deliver physical items.

This is actually why Bezos has effectively started a new company, or seeded a new company, called Prometheus, and it is supposed to be an AI company. Basically, the idea of Prometheus is that—I’m sure you all have heard the story, the famous Greek story—Prometheus is the man who ended up bringing fire down from the heavens, from Mount Olympus, to the average person and giving humanity the ability to grow and create.

The gods punished him massively for it. They actually chained him to a mountain and had an eagle eat his liver every day because his liver would regrow. But, long story short, what Prometheus does is try to lower the cost for the production of physical goods. This has always been Jeff Bezos’s approach, and I think that obviously is going to be quite valuable.

No, no, I’ve got one last thought I have to share, guys. So, back to Apple—the spectrum between Apple and Google, with Meta in between.

AI—the US used to be ahead in software. The internet was our thing, and we outsourced atoms to other countries where labor was cheaper. Now we’re sort of behind China, but we’re not behind most other countries in that regard. AI has basically commoditized software, leaving the US—the tide has gone out, and the US isn’t really wearing any clothes there.

Obviously, now there’s this big race to capitalize on atoms. In Walter Isaacson’s biography “Elon Musk,” which I recommend everybody read or listen to, Elon Musk has this thing in his companies called the idiot index. The idiot index is: What is the cost to buy a product to put into my spaceship or Optimus robot or Tesla, like an actuator or a pipe connector or a chip or anything? What’s the difference between what it costs to make that product and what it costs me to buy that product?

It’s the sum cost of the raw materials, the input commodities, versus the output price. That’s the idiot index. Elon is willing to buy parts that have a low idiot index, and he makes parts himself where the purchase price reflects a high idiot index.

Basically, Jeff is trying to bring down the idiot index for American parts because that’s where we’ve fallen behind. And it’s technically our national security’s biggest problem. America’s national interest is most exposed as software gets commoditized because there’s nothing to cushion P&L anymore here—nothing to cushion the GDP without that sort of prowess in the world of atoms.

That’s Travis’s thesis, too. I think the way that you express this in your trading and your daily life is by exposing yourself to atoms as much as you can and distancing yourself as much as you can from knowledge-based outcomes.

Google, jury’s out. Apple has basically a monopoly on the atoms in every rich person’s pocket in the entire world. They’re doing nothing with it, which is freaking crazy, in my opinion. Meta is in this weird middle ground between humans, sort of IP, and atoms. They’re not a lost cause, but I think they have some real soul-searching to do at the top of that company. I’m hoping for bold action from Zuck before I try to reinvest in that one.

I think that’s as good a place as any to wrap it for the day.

I promised the people in there that I would talk briefly about Bitcoin and custody risk. There have been some hacks on Bitcoin recently involving multisig wallets. There’s some structural risk.

Basically, my opinion is that there was a firmware bug with one particular type of cold wallet that resulted in a drain. I think if you go with one of the majors, Ledger or Trezor, and you self-custody, and you keep your seed phrase in a bank vault or someplace safe, you're good. Or in your head, best of all. I really still believe: not your keys, not your crypto. Just self-custody.

Don't use a wallet with lots of software on top of it. Use one of the OG wallets. I also think you're fine storing your crypto on Kraken with a passkey. They're probably proven enough at this point to be trusted with your funds if you can't self-custody without losing your seed phrase or if you don't have a bank vault.

I think Bitcoin's going to be in no-man's-land until this AI trade cools off or rates come down. Rates probably won't come down in the near future. So, basically, in response to the thing that I said I would talk about on this podcast as a promise to the chat room, focus on what we talked about on this podcast. If you're capable of self-custodying with one of the OG wallets, do it. In terms of trading Bitcoin, stash it for a decade and don't think about it right now. That's my response.

Look, I have a slightly different take on this, which is that for the vast majority of people out there, unless you're technical, because it actually does still take some level of familiarity, just hold your crypto on Coinbase. I mean, no, or Kraken. I wouldn't even say a U.S.-based exchange, because I don't know if I trust Gemini or these other exchanges, but basically I think if you hold it on Coinbase, you're kind of mostly fine.

If you want, you can teach yourself how to use one. I wouldn't use a Trezor; I'd literally just use a Ledger, too, because I think there have been some issues with Trezor in the past, and I think Ledger just has better people working on it. You can try to teach yourself to use it with a small amount, and then if you really feel comfortable with it, do it.

But I've also heard horror stories with Ledger, and if you're not somebody who is highly detail-oriented, it can be very easy to lose your grip. Regardless of whether you get hacked or not, there are so many other ways to mess it up, from losing your seed phrase to losing the actual Ledger itself. It's just a problem. So my take is that the average person is probably better off holding it on some sort of platform.

Like, you know, Robinhood.

Robinhood, because Robinhood, by the way, if Robinhood ever gets hacked, they are for sure going to have to reimburse people.

Enable multifactor authentication, though, if you store it on Robinhood or Coinbase. Do the 2FA.

For sure.

With that, it is Friday. So, to all the Jews and non-Jews out there, Shabbat Shalom. I hope you have a—

Shabbat Shalom. I hope you have a wonderful and restful weekend.

Take some time off and just relax. Great to see you, Avi.

Great to see you, too. Take care, Jonah.