《Softwar：Larry Ellison 与 Oracle $ORCL 的亲密画像》（Fintwit Book Club，2025年7月）

You're about to listen to the Yet Another Value Podcast with your host, Andrew Walker. Today's episode is my monthly book club with my friend Byrne Hobart from The Diff. We read Soft War. It is an intimate portrait of Larry Ellison from Oracle fame, as the book calls it. It was published in 2003, and I think we both really enjoyed it. I'll be honest, I don't know if I would recommend the book, but I think we both really found it thought-provoking, just seeing inside how tech migrations go, seeing inside the company, and thinking, using the benefit of what we know from the past 20 years, about how Larry Ellison, Oracle, tech, and everything evolved. I think we both found it really interesting. If you do read it, again, I think you will learn and find a lot thought-provoking. Skip all the sections on the boats. That's the only thing I would advise you, and you'll have a much better time reading it. But it's a really fun podcast. I think you're really going to enjoy it. We'll get there in one second, but first a word from our sponsors.

Today's episode is brought to you by FinTool. Fin is the AI junior analyst tailored specifically for individual investors. Everyone in finance is racing to figure out how AI can best be integrated into their investment process. And one of the biggest areas that is catching on with institutional investors is analyzing SEC filings and earnings call transcripts. FinTool takes hours of combing through filings and earnings call transcripts down to seconds. Whether it's comparing the current call with prior quarters, finding that sneaky change in the footnotes, or compiling the key facts into an easy-to-digest one-pager, FinTool is saving you hours so that you can go deeper and search wider because your time is better spent turning over more rocks or researching the things that AI can't. Go to fintool.com to transform your research process. That's fintool.com.

It's going great.

Quick disclaimer, nothing on this podcast is investing advice. You can hear the full disclaimer at the end of the podcast. All right. Hello and welcome to the Yet Another Value Podcast. I'm your host, Andrew Walker. It's our monthly book club. With me today, I'm happy to have my co-host book club partner Byrne Hobart from The Diff. Byrne, today we both read Soft War. It's Larry Ellison. I guess it calls itself an intimate portrayal. It was released in 2003, which I think makes it a really interesting retrospect, because the bulk of the book focuses on Oracle right during the dot-com crisis and right at the end of the crisis. I want to talk about all sorts of things here because it was really interesting reading this with 20-plus years of hindsight bias, knowing Larry Ellison, and knowing where the future of the tech world and everything else was going.

Byrne, this month we both read Soft War: An Intimate Portrait of Larry Ellison and Oracle. It was released in 2003, which I think makes it a really interesting retrospect, because the bulk of the book focuses on Oracle right during the dot-com crisis and right at the end of the crisis. I want to talk about all sorts of things here because it was really interesting reading this with 20-plus years of hindsight bias, knowing Larry Ellison, and knowing where the future of the tech world and everything else was going.

You and I both read this book. It's 700 pages, and 250 pages of it is probably devoted to Larry Ellison telling boating stories.

I will confess I did not read the boating parts very attentively. I'm glad he enjoys his boats. It seems like they go fast—not quite fast enough sometimes—but that was not my favorite part.

For those who probably haven't read it, the middle 150 pages is Larry Ellison describing boat races he's been in for 150 pages. I'm with you. I read the first 10 pages and was like, “I'm out.” I'm skipping it.

But I want to ask you: If you had read this book in 2003—let's start there—you read it in 2003 as a stock analyst. What would your opinion of Oracle have been? If you could fast-forward 20 years and see what happened, what would you have thought about Oracle?

Oh, you mean having read it, yeah.

If you read this book and somebody puts a gun to your head and says, “Buy, go long, or short,” what are you going to say? Yeah, that is such a tough one because I had such mixed feelings about Oracle and about Ellison. I noticed this weird thing where he likes to complain about other people's flaws. He has people he likes and people he hates, and he'll tell you why. Everyone he doesn't like is actually more like him than the reference category he likes.

He talks about how Oracle is an engineers' company, that they care about technology, and that there's a reason all the successful tech companies are run by programmers. But you don't have any scenes in the book where Larry says, “Okay, I'm going to look at the man page and debug this shell script that I wrote.” You only have scenes where he's selling. He talks about how he doesn't understand the salespeople, but then the next scene will be him pitching a bunch of CEOs and having this incredibly effective deflection when they say, “We've spent tens of millions of dollars on your products, and we've gotten no return from it.”

He talks about how he can't stand that everyone in the industry is dishonest and that they're all making up fake benchmarks and things. I just feel like Oracle would have a much higher market cap if every benchmark that I've ever seen in an Oracle ad were, in fact, true. They show some pretty incredible numbers that I assume are true in some context, but not for everything.

So I feel like, looking at that, I would have said Oracle is as technical as it needs to be. They clearly do have to have really good software engineers. I think if I'd looked at it at the time, maybe I would have refused to go long for completely incorrect reasons. I could have said, “It is absurd that he has this thesis that we're going to run everything on distributed systems with commodity hardware, and we're going to let these crazy amateurs who wrote their own version of Unix because they have too much free time—they're weird nerds. Would you actually put anything serious on a platform like that? Of course not. You want to buy something from Sun instead.”

Then, when Larry is going on and on about how the future of computing is a $500 smallish, mobile-ish device, he still thinks that it's a laptop, but he's basically describing what the smartphone evolved into. A pretty simple device, pretty cheap. Everyone has them, and everyone uses them all the time. They are basically a browser.

When you use mobile apps, what you are basically doing—unless you're playing a game—is using a topic-specific browser formatted for the specific thing you're trying to do. LinkedIn on an app is very much LinkedIn in the browser. The Interactive Brokers app is the same kind of interface—not the same interface, but the same information and the same capabilities as the site, et cetera.

So he got a lot of stuff surprisingly right. By the end of the book, I was just kind of confused. I really could not tell if Oracle was a good company selling products in a hard category to do well in, or if Oracle was just really good at selling and the technology was okay, but they would always overpromise, so you would always be somewhat disappointed.

I think you can read it and feel like this is just a catastrophe in the making—that the company nearly went under one time, and that would probably happen again—and that obviously would have been completely wrong. But there was also this striking thing as I was reading, which I think is part of your question: A lot of the details were worth reading in part because they were things that I took as a given, but that somebody actually had to articulate and ship.

Stuff like, yes, everything happens on a server, or you're interacting with it through a browser. That was something that people like Ellison had to say was going to be the way things worked.

But then other parts of it just seemed completely timeless. People have these big, splashy projects. They do ERP transitions, and they don't really work well; there are incompatibilities and mistakes. Just last year, Lamb Weston had this big earnings blowup because of an ERP thing, which I thought was particularly funny because so many of the hedge funds that hold Lamb Weston were software-focused funds.

They just happened to own a potato stock. That's why I first heard about it: I was looking at 13Fs and wondering why so many growth funds were buying potatoes. If those software-focused investors didn't sell at the right time, they lost money on potatoes because of software.

That part felt kind of timeless. Then you look at the cover and you look at Larry Ellison today, and you're like, “Okay, he basically looks the same.” Incredible.

He does talk a lot about aging and longevity research at the end, which was interesting. No, look, you hit on so many things I wanted to touch on.

Just to quickly touch on the Lamb Weston thing, that was one of the things I really took away from this book. It is a common trope among investors: A company is implementing an ERP transition, so short the stock. When you read this book, I've never been inside a big company undergoing a tech transfer—maybe I was when I was at a big company, but I've certainly never led one. You really start to understand why you would want to short a stock.

To you and me, I'm sure it just seems like, “Oh, you upgrade some servers.” It is much harder than that, and you kind of get that in the book. But—

Yeah, it looks like a partial brain transplant where we're going to keep everything running, but we're actually snipping out a little bit of the brain stem.

We're going to pop in a new, faster brain stem, and we're pretty sure this will work fine and you won't have a seizure and die.

They've got 2 really interesting stories in there. One is when Oracle upgrades to its own product, and they say that for 10 days the company is basically shut down. They're like, "Thank God we decided to do this over the Christmas holidays, because we can't do anything." You're like, that's a tech company that basically shuts down.

And there's another where GE Power—I think it's in Hungary, I want to say—is upgrading. They try to bolt their own systems onto Oracle, and they're like, "Oh, my God. If we were doing this for real, everyone in this plant would have just died," or something. So they're like, "Oh, it's so difficult. We have to use the software and adapt our procedures to the software."

I thought those were really interesting, and it explained to me why there's a trope of ERP shorts. It is really hard just to stick with. I want to talk about all the things in the future, because you hit on that and it's so interesting when you read this, but let's just stick with this: would you have gone long or short the stock if you had just read this book?

And look, when people say "short," I think you mean short on a relative-value basis, right? So would you have bought the QQQ and shorted Oracle?

Yeah, because I'm not talking unabashed alpha. Go ahead.

Right. Yeah, I do feel like I would have bet on underperformance.

I would have, too. The reasons I would have given you were, one, a lot of the stuff in the early 1990s. They almost die in the early 1990s, and I do want to talk about that later. But from the early 1990s until the end of the book, Larry's always complaining. Everyone's complaining: "Oh, Oracle, all of our sales..." It's hard to manage this business because all of our sales happen on the last day of the quarter.

Basically, they're constantly having issues with channel stuffing. You're like, "Okay, well, that's one." Two, the CEO gets this book published, and literally, the thing he's most passionate about—and the thing that the book spends the most time talking about—is probably his boat racing. Oracle sponsors the boat racing. All of the analysts are worried about literally how much time he'll take off next year to go race boats and stuff. He almost dies racing boats. So you have a CEO who's distracted.

Then there are lots of related-party transactions, starting from when Ellison founds Oracle while he's working for—I think it's OMX. He's working for them, and he puts out a bid and has his own company, Oracle, that he founded, take the bid. Even though Larry isn't doing any of the actual programming work, he takes the majority of the shares.

Then throughout the book, there'll be, "Oh, yeah, Larry's got 100 million in this company, and they're making a product, and Oracle's partnering with them," or Oracle thinks they should buy them. So you've got related-party dealings, channel stuffing, an almost-dead, distracted CEO. That's a trifecta, or that's a quadruple there.

Yeah, it felt like the kind of company where you just don't know when it will blow up, but it's inevitable that it's going to blow up. Then it talks about Larry's habit of finding someone who seems really great, promoting them, deciding he promoted them too fast, and now he's mad at them because he's overpaying them.

He also blames them for whatever. It seems like—I forget the guy's name—but there's one guy where Ellison basically gives him the mandate: "Okay, just grow revenue. You're a sales guy, you're a deal guy. Do deals."

This is the Mormon, I think. There's a Mormon who they're like, "He's the greatest salesperson. Just sell, sell." Then when he sells and sells, they get mad at him for selling too much.

Yeah, and for margins. That also felt kind of petty. Do you really want to fixate on margin when your company's growing really fast and it's growing into something that may or may not be the standard?

Ellison's constantly talking about how you need standards, you need interoperability. It feels like there are 5 different cases where he uses the analogy of having a phone system, and that's like your software setup. If you have one person who speaks French and one person who speaks Japanese and they try to have a conversation, you realize they're mutually unintelligible. So you need this common language, and it's the Oracle database that's under the hood for everything.

He's talking all the time about this standardization effort, which basically implies that every incremental point of market share is worth more than the previous one. You would think that would be a time when, if salespeople are upgrading themselves to business-class seats when they go on flights to visit customers, that's probably not the biggest problem for the company to focus on.

You said petty, and one of the things the book has the most of, which surprised me, was how much backstabbing there is when it comes to management. The thing it has the most of is Larry getting sick of his right-hand man at the time, having a falling-out, and then firing him right before his stock options vest. I think it happens with 5 different executives.

I'm not on the inside of multibillion-dollar tech companies, so maybe this is common, but I was shocked by how often you'd have a high-up person who gets fired by Larry right before the options vest. Larry will basically say, "Hey, if I fired my secretary, she would get 2 weeks' pay or something. Why should a highly paid professional, when you fire them, be entitled to multimillion-dollar severance packages?"

On the one hand, I get it, but on the other hand, it's just so common throughout the book. He's firing people right before their options vest. Again, if I were throwing in red flags, I'd be like, "Hey, always firing your right hand, always having falling-outs with your right-hand man, and firing them right before their options vest." This is some really strange behavior. There's a fifth and sixth red flag. Anything there you want to add on it?

No, I thought that was all pretty aggressive stuff. On the red-flags thing, one of the funnier bits to me was when he's talking about the early stuff and he casually mentions that Oracle is named after a CIA project. Then they land one spy agency—it's like the Defense Intelligence Agency or something—as a customer, and that opens doors for every other intelligence agency.

It did feel like you could definitely spin some interesting conspiracy theories. It's sort of like, if you've ever seen people obsess about this with Facebook years ago, about how it was funded by In-Q-Tel and was probably this big intelligence project. Early Oracle kind of has that vibe, where you can definitely tell a story: this guy works at the CIA, then he creates an independent company, and all of its revenue is from different spy agencies. And what does it do? It stores everyone's information about everything.

The problem with that as a conspiracy theory is that it's a very post-telco-buildout conspiracy. If you have a back door into some software, you can actually access it and exfiltrate information, whereas in practical terms, even if there's a secret password to the database or something, you have to physically get to the machine or get access to the machine. It's not as easy as it is now.

So I thought that was fun. I'd never heard conspiracy theories about Oracle as an intelligence front or as a former intelligence front, but it's just one more thing that makes it look a little bit weird.

As you say, at the end of the book, September 11 happens toward the end, and Larry Ellison goes on this big kick where he's saying, "Hey, we should have"—I believe it's government ID cards, like a government database. Everyone should have government IDs for security purposes.

What would a government plant say now? This would hugely benefit Oracle's business and stuff, but what would a government plant say? A high-powered CEO going off and saying, "Hey, everybody needs a government-mandated ID card." That sounds very, very close to the government line.

Let me go look for it. Look, you and I are talking in late 2025. It's hard to read this book without bringing our knowledge of everything that comes after it. I think you've already hit on some of Ellison. A lot of what he's talking about—and you don't know how much of it is us projecting onto it. You read a science-fiction writer in the 1960s, they get some stuff right, they get some stuff wrong, and a lot of people celebrate the stuff they get right.

I don't know how much of what Ellison's talking about is us giving him credit for the stuff that comes true, or how obvious it was. Whatever. The things that were really striking to me—let's start on the personal side. Larry Ellison becomes really close with Steve Jobs at the end of the book, and he's always very close with him. Then later he becomes very close with Elon Musk.

It's hard for me to read this book, particularly without seeing all the shades of Elon Musk coming through. That goes from the relationships with women, to the salesmanship, to all sorts of things. I'd love to start by talking about this: were you reading this book kind of like I was, with a little bit of the Elon Musk to come?

You know, I hadn't really thought about the Elon parallel. It’s definitely there now that you mention it. Elon’s vibe is partly that he is very technical and, from what I understand, he is actually really technical and really smart. But his public persona is very much a sales persona, and Elon and Larry both like fast cars and expensive real estate sometimes, although it feels like Larry enjoys his more than Elon enjoys his. So they do have some commonalities.

But there’s also the fact that Ellison just deeply identifies with Oracle. You can read him saying that he’s going to quit at any time and focus on molecular biology, and that’s what his real passion is. But I feel like—even though I read the book around when it came out—I remembered almost nothing from it except the footnote. We should talk about the style, but the hot-pink tank-top footnote was my one firm memory.

What’s the hot-pink tank-top footnote? I can’t remember.

Oh, okay, we can do a sidebar on this. The structure of the book is that Ellison is technical—he describes himself as technical—but you constantly see him doing deals, and you never see him writing code or proving theorems or something.

The book opens by discussing the negotiation with Ellison over what the book will be. You could do, “I’m going to ghostwrite your autobiography.” You could do, “I’m going to write a totally unauthorized biography, and I’m going to talk to everyone and dig up all the dirt.” Or we could collaborate a little bit about how we’re going to collaborate.

The author’s agent comes up with this idea: Why don’t we have the author write the manuscript and give Larry unlimited flexibility to put in whatever footnotes he wants about whatever point he wants to make? A lot of them are in chapters about a falling-out with some former executive. Basically, every time there’s a quote from that executive talking about what it was like to work with Larry or how he felt about getting fired by Larry, Larry says, “This guy’s full of it. This guy sucks.” It’s very petty, to use the petty word again.

But there’s—so let me see if I can find it, because it was a great little detail. It’s talking about a meeting that took place on Ellison’s boat, and he—let’s—okay, here we go:

“I remember going over to Larry’s house in Atherton. This was my first experience with Larry. Larry shows up in shorts and a pink tank top holding a glass of carrot juice. And I’m thinking, ‘This is my kind of guy.’”

And there’s a footnote. Ellison writes, “I met Matt out on my deck just as I arrived back home from the gym. I was wearing a black cotton tank top with a logo on it. I have never owned nor would I ever wear a pink tank top. That is very important.”

I might have skipped that footnote because I don’t remember that one. The footnotes throughout are quite funny, just because he gets accused of so much, and every time he basically says, “I know you are, but what am I?”

Yeah. It was like turning the Shaggy song “It Wasn’t Me” into a full book. Every time Larry’s accused of doing something bad, it wasn’t me.

Then he compares himself to Galileo at one point. He says it’s like Galileo and the Pope—

The Pope telling Galileo to stop researching or something. Yes, I remember this. So Tom Siebel, at one point—I like this term—he says—

“You’ve fallen for the lies of Larry.” I think he accuses the author, but he might just accuse the press of falling for the lies of Larry, right? He says, “Look, Larry lies and distorts everything.” The author even says the press has come to believe that he exaggerates everything.

Again, I think Larry’s a little more humble when it comes to the boating stuff. Basically, the author and Larry are both saying that Larry might not be the best boat captain in the world or whatever it is, but he’s professional-tier.

I was struck by the lies of Larry again. If I bring it back to Elon Musk, there’s been the recent scandal with Elon Musk and his Diablo stats. He basically says, “I’m the best gamer in the entire world” at two or three really popular games, and the really popular gamers are like, “Absolutely not.” It just struck me that the lies of Larry run through the book—the salesman.

I’d love to talk about that as it relates to Larry and Elon, and then as it relates to Larry and his subordinates. It’s really interesting how so many of Larry’s subordinates get fired, he hates them, and then they go on to run big, successful companies. Tom Siebel would probably be the big example.

Then, if we can bring the benefit of the future to it, the book closes with Oracle making the hostile bid for PeopleSoft, which I want to talk about. Tom Siebel—Larry, I think, says that if he had 1 bullet and you and your dog were there, it wouldn’t be your dog he was worried about. Siebel accuses Larry of being willing to kill people’s dogs to sell products, and then Siebel shows up at an event with his dog, both of them wearing bulletproof armor, which I think is 3 years after this book was published. Oracle buys Siebel for about $6 billion. So it’s hard to process. Anyway, I threw so much at you. I’ll just toss it over to you, and we can talk about any piece of it.

Yeah, there’s a lot to go through. The rivalries were really interesting. One of the things that I thought about periodically while reading this book was that he and Benioff seemed not to hate each other as much as any former protégé. And I can’t tell—

I forgot Benioff was a former protégé as well. It may be that Benioff actually noticed this pattern and just decided at some point, as the next vesting date was coming up, that either they were going to be enemies forever or he was going to leave Oracle and they would stay friends, and chose what turned out to be the right call.

The drama around how much they competed, how aggressively Oracle competed, and how they sort of buried the hatchet—I do feel like the Siebel acquisition was probably more of a victory lap, with Oracle kind of solidifying how dominant it was.

Let me pause you on the PR point, because this comes back to the red flags we were talking about with Oracle. I’ve read a lot of business books. I’m sure you’ve read a lot of business books. I’ve never read one that had so many mentions—aside from 1, which I’ll say in a second—of quarterly earnings, stock-price response to earnings, and the company really caring about quarterly earnings. The only other one that I can remember is Enron. This is the only one I can remember.

Again, this is 20 years old, and Oracle is clearly not a fraud. I can feel comfortable saying that. When I saw the channel stuffing in the 90s and then I saw, in 2000—I mean, the dot-com bubble—they were talking about it in every quarterly report. I believe there’s one at the beginning of the book where Oracle’s stock goes down because they miss earnings by a couple of cents or something, and they take out ads in the Financial Times to describe why their business is doing well and why Wall Street was wrong.

I wasn’t sure: Was that a sign of the culture? Was that a sign of the times? Because, again, we’re right on the heels of the dot-com bubble, and Oracle, for a while, holds in there. Or was that actually a business strategy where they were saying, “Look, these companies are comparing Oracle—which is still kind of a startup until maybe the late 90s—with buying IBM”? It’s important that they know this is mission-critical stuff. It’s important that they know we’re a strong business. We’ve got momentum. People are still switching over to us, even if the stock price has gone down; we’re still taking share.

You want to be with the new players. So how did you read that?

Yeah, I think maybe that social proof thing is really important. If you are an enterprise software customer and you've heard of this company, you've seen them on CNBC, and the stock chart is going up, you feel a little bit comfortable that this is working. If you are hearing about them in a negative context, where they're missing earnings and the stock is in the dumps, then maybe you do feel like they are has-beens and the market is telling you that there's something new to think about.

Was that an early-2000s thing because people were still getting used to the IT budget, or do you think that's still the case today?

Yes, if your supplier is in absolute distress, you're going to be scared. But no, today everybody understands, and you might be willing to take a small startup or something.

Do you think it's changed, or do you think that was an early thing?

I think the bigger the contract and the bigger the company, the more that kind of social proof matters. This is actually one of the ways that VCs in that space can add actual alpha and deliver some kind of value other than just the cash that they provide. If the VC's logo is on the company's page, that at least tells you that Sequoia does not think this company is going to vanish anytime soon, so you can be a little bit more willing to work with them.

I think that part probably does hold. Maybe the fact that Oracle was fixated on its stock price is also partly evidence that they did actually care about their software engineers, because the salespeople are getting paid commission. It's still variable, but it's realized right away, and the salespeople all understand this. With the engineers, certainly the people who joined early and got options and did well from that, they recognize that equity compensation is a really powerful way to earn a lot of money from working at companies that do well.

Once that's the case, if the narrative around the office goes from, “I joined Oracle 5 years ago and I just bought a house for cash, and I still have $1 million of liquid assets, so I'm doing great,” because it's the late ’90s, to, “I joined Oracle and I'm actually poorer now because I held all of my stock and bought more, and the base pay has not made up for that,” that is pretty bad for morale. If you're a salesperson and you're just complaining that your commissions are down from last year, your sales bros will just razz you and tell you that this is a personal failing. But if it's the stock price, it's something a little different.

I think there are probably a bunch of different forces there. The retail investors were not necessarily a bigger force than they are now, but I think they were more of a normal force. There were retail investors who would actually buy things like AT&T and collect dividend checks, and that was their thing. The stock had a lot more salience, and running ads like that was probably more understandable in that context than it would be today.

You're a meme stock; you're trying to become a meme stock. No, the other thing that was interesting was the sign of the times, obviously, but how much press sell-side analysts get in here. If you read something starting in the early 2010s through today, I don't think sell-side analysts would ever get any press in a tech book or anything. I can't even imagine in what context they would get press.

Well, yeah. If I were writing a book about Oracle today, what I would do is use sell-side quotes from sell-side reports as just a snapshot of what the consensus was at that time. I think that is part of the social function of sell-side research. But when markets were less efficient, it actually made some amount of sense to pay an analyst to research a company and figure out if it was a buy.

I did think it was interesting because this is still true: the Gartner analysts were so important.

I was wondering about that too. I don't think they carry the same weight anymore. Gartner, Forrester, all of that—Ellison hates them, and when they make recommendations, people are really pulling budgets based on them.

Yeah, he has, I think, in one of his footnotes, he's comparing Ellison with one of them. I forget who was at Forrester and who was at Gartner, but he says that with one of them, he will cite sources for everything. You can always double-check his work, and you can make sure everything adds up. If he's wrong, he'll correct himself. The other one has this amazing technique where she just makes up everything, and then you can't check any of it.

Let me switch. You and I bonded over this 2 or 3 years ago: the companies that lost the most market cap value in a day. Facebook did it at one time, and a couple of other companies did it as well. I think what was interesting was that if you had just bought the companies that lost the most value in a day, you actually did pretty well as a portfolio.

There was one line in here that Oracle, at one point—I think it was during the early ’90s, during one of its biggest drops—had the biggest market cap drop in 1 day. I was like, “Boom, it's another example. Just buy the companies that have the biggest market cap drop in a day.” Okay.

At that point, I think if you do that—well, actually, you are probably buying large-cap growth, right? Owning large-cap growth has been a pretty good thing to do over much of financial history since then. To some extent, there is a little bit of free riding on that factor exposure.

On the other hand, it does make some logical sense. If you're big enough to lose a lot of market cap, it means you're a big company. If you're actually in a position where you can lose a lot of market cap, it's either that you had some kind of huge accounting scandal, or a huge but nonfatal accounting scandal, or your business is actually uncertain enough that, even at that massive scale, there are a lot of open questions about where it can go and how big it can get.

That at least gives you something. If nothing else, if you wrote a business history of the United States where you just took each of those big drawdowns and looked at what led up to them and what the consequences were afterward, you would probably get a nice summary of the highest-impact companies. Those are the companies where they can be very big and you can still have a huge range of estimates for what they're actually worth.

If you look at Google today, what number do you put on Waymo? Is Waymo an interesting cash sink that is good for the brand and may eventually turn into a real business? Is Waymo actually a lot of the value of the company because they have kind of cracked the code and can scale? You can still debate a lot of that stuff about big tech companies today. It is a signal that there are a lot of different interesting futures.

Let me switch to, speaking of Google and Oracle, when you read this book, a lot of the Oracle leaders who leave—often, as we discussed, on contentious terms with Larry, but some, like Marc Benioff, on nice terms—become literally the future of tech, right? Benioff runs Salesforce, Siebel ran Siebel, and several other Oracle spinoffs, even though they're not necessarily Oracle spinoffs. Their executives are getting recruited left and right to go run big, tech-focused companies.

My question is, when you have some companies where the pipeline is full of CEOs—you know, famously, GE would be the example—Oracle is a great one. You can probably list 10 Fortune 500 executives who came out of Oracle. Google obviously has lots of executives come out of it. Facebook has a lot of executives come out of Google, though interestingly, it tends to be Facebook executives who take their billions, found their own company, and then kind of find their way into a billion-dollar company, versus other companies hiring Facebook executives.

But there are some companies where you don't get a lot. I would point to Apple. I can't think of a single ex-Apple employee who's at the top of a really large company. I'm probably missing some, but Apple famously doesn't produce many. Tesla and all of Elon Musk's companies, obviously, are another example.

Yeah, Thomas Kurian has a cameo at the beginning and then at the end. I like cases where a book will just randomly mention someone offhand who ends up being famous later. There was a book that was a history of Lazard, and it mentions a random Lazard associate, Mark Pincus, who had a year or two there.

I like those little cameos, but a lot of them do end up being in technical but customer-facing, salesy kinds of roles. There are some companies where the pipeline is full of CEOs, but others where you don't necessarily see that same pattern.

For the most part, I can't think of a single high-level executive who's spun out of the Tesla sphere. Tesla, Elon Musk, Oracle, Ellison—it’s interesting because you've got these 2 guys who are connected, and I see a lot of similarities in their styles and preferences. One company is just spinning off executives left and right; another company, to my knowledge, is not spinning off any.

What do you think it is about one company that creates a management and CEO spin-off factory, while another company—I don't want to say the talent's poorer, because I've heard great things about Tesla employees—isn't producing high-level talent that's going elsewhere?

Yeah. I think there are a couple of phenomena. I wrote a piece a long time ago about business mafias and how there are different companies where you hear about them.

PayPal, famously.

One of the things that I noticed as a commonality is that if you take the core part of that company and look at how many people it is, it's usually relatively small. For PayPal, the core was the software and the dealmaking side, excluding the customer-service piece, which was in a separate office. You had around 200 or 250 employees, something like that. Tiger Management was actually in a similar range for its full team.

I forget who else I included in this list, but one of the common patterns is that to get a really good business mafia, what you need is a company where they have a premature exit. In the case of Tiger Management, they made some bad macro bets at a time when they were at peak AUM, and it just fed on itself until they wound down.

You have a bunch of people at that company who still have a lot to prove. They have enough money that they can try something else; they don't have to immediately get a job because they can't pay rent. They all know each other, too. Part of the theory that I had on this is that if you have this small set of people and they're all working insanely hard, their entire social circle is saturated with connections within that firm. They all know each other, so they're part of this narrow cohort.

I think that's one piece. For GE, I think it's just a different dynamic. If you run a company that's fairly decentralized, with all these different divisions doing different things, you can have a lot of people who are doing many CEO jobs.

This is one reason Palantir has produced so many founders, despite being a lot younger than these other companies. A forward-deployed software engineer is basically running a sort of temporary company within another company and making a lot of CEO-ish decisions. They're balancing technical constraints against business constraints, working within their organization and within other organizations, and trying to figure out where you just pay the opex and where you use capex to eliminate some cost. They're making a lot of that kind of executive decision.

When you read about GE—which, incidentally, I thought was going to be the company you'd mention as a '90s company where books about them are like, every chapter could be, “Here's how they hit their number that quarter”—

You know, I read the Welch book a while ago. Maybe I wasn't focused on it, but I do remember that was one of his things. I don't remember them taking out advertisements to rebut that, but yeah, go ahead.

So they would do stuff. I think there was some anecdote about Immelt—it was either— I think it was Immelt actually doing intercompany deals on the last day of the quarter, when he ran plastics or something.

On the last day of June, he's asking some other GE subsidiary, “Can you please put in a purchase order for $5 million worth of our plastics? We'll buy something from you next quarter. We'll make it up to you somehow.”

If they have that culture where it's fairly decentralized, where you have your own P&L, then you'll have a lot of people who could run a company. Presumably, a lot of those people join GE because they want to get to the top.

If you have this sort of GE management culture where everyone understands that they have these shared experiences, and everyone feels like it's a pyramid with a lot of attrition at every level, then at the top you have, “Before it was Jack Welch, and soon it's going to be me.” You probably have a lot of people who feel really salty every time there are 5 people who all expect a promotion, 1 of them gets it, and the other 4 have to go run AlliedSignal or something.

Look, that happens. GE—that's what happened, right? Immelt gets promoted behind Welch. It’s Welch, then Immelt, and the other 2 top guys who were in contention for that leave. One goes and becomes the CEO of Home Depot.

Let me ask you a question on the PayPal mafia thing. Your contention was that one of the reasons you get one of these mafia-style things—where you've got a company where there’s an early exit—is that PayPal had an early exit. What basically happens is that you have a company crushing its category, with super-successful, super-ambitious people, and probably a bunch of late-20s and early-30s people who would have stayed at the company, gotten promoted within the company, stayed there, vested, and all that. Instead, they get taken out.

They probably have a lot of money in their 20s and 30s, they're all connected to each other, they're super ambitious, and now they've got a lot of money. So they're seeding and starting each other's funds, VC firms, next ventures, and all that sort of stuff. Is that kind of what you're thinking behind that?

Yeah. Exactly.

They also had a practice of hiring through their networks. They would always ask people to name their 5 smartest friends and see if they wanted to work for PayPal, too. So you had some pre-existing connections that then got a lot stronger.

Oracle is kind of a mystery. I guess, to take a step back, you can look at those mafias, and they're a really interesting phenomenon. But if you look at companies that don't have as much of one of those—like, it took Google a long time to have that, and it just doesn't have the same kind of network density because people gradually leave and start other things—the people involved may create more value in aggregate.

I don't know if this is still true, but the last time I looked at this, more market cap was created by Google not having a Google mafia but simply by having Google than—

That's why I think the PayPal example was interesting: the early exit. The contention here would be that Amazon probably would have had a lot more, except people stayed at Amazon. Amazon just kept going successfully, and instead of exiting in their early 30s and being like, “I've got nothing to do,” they stayed for the next 7 years.

They were $100 millionaires at Amazon. They retired, or they stayed at Amazon, whatever. Now they've got a family, and all their connections are still at Amazon. They don't have that network density.

That's why I was so interested, and that's why I do think Oracle is an interesting example, because they do still have that kind of mafia despite not having the early exit. Maybe that's because Larry Ellison's exiting everyone early, and their options and stuff.

You want to know another thing? Ellison's always trying to hire the most talented people. Are you surprised that Ellison hasn't been canceled? He says, “I didn't get why we would hire the brightest, most up-and-coming young salesmen”—and in the '80s and '90s, they were mainly men at the time. “I didn't get why we'd do that. And for our EAs, we'd hire people without college degrees.”

So they start hiring Stanford grads for all of their spots, basically. They're also hiring super-cute Stanford grads, and Ellison's just dating every one of them. I'm a little surprised it hasn't come back to bite him. He's got a playboy reputation.

Yeah. I think it's for the same reason that there aren't good Oracle conspiracy theories: people don't realize how much Oracle software they're using because they're using it so indirectly.

Whereas, if something is customer-facing, it is a big deal if the CEO is chasing the assistants and things like that. I think that's actually a big part of it: there's just a lot more tolerance for misbehavior because you don't even know what to boycott if you're boycotting Oracle.

Either you say, “Okay, I am not going to buy an ERP system to manage my home or whatever, so take that, Larry Ellison,” or it's like, “Okay, I am not going to work with any Fortune 500 company because they are all tainted by the Oracle association, or their suppliers are.”

No, I hear you. Obviously, I'm not accusing anyone, but it strikes me that the Astronomer CEO having that viral moment could very well have been Larry Ellison. That was 2 very high-up people in Astronomer. It very well could have been Larry Ellison and an EA, because one of the stories that comes out—I mean, there's this crazy story with one of his former girlfriends, who's an EA, suing him and all this sort of stuff, and criminal charges.

It's crazy, but one of his defenses is that he has another EA in a different country, whom he's dating, come to his defense in the story.

It’s like the man was dating 2 EAs in 2 different companies inside the same firm. Let’s get—let me switch to the end of the book. I found the end of the book to be really interesting, so the afterword is PeopleSoft, which I don’t think we have time to talk about. That becomes a landmark DOJ case, which I think is kind of interesting, because throughout the book he’s hating on Microsoft and their DOJ case, and then he gets involved in what?

Yes. But do you want to say anything on that?

No, just like it’s another instance of Larry saying, “I can’t stand these smarmy salespeople,” and then being a smarmy salesman, or saying, “We’re really an engineering-focused company, and we would never lie about stuff,” and then lying about stuff. So it was another instance of that where he definitely hates people who remind him of his own worst traits.

But the reason I mentioned the end of the book is I found it really interesting. The CEO of Oracle today—so, this is a benefit of hindsight—is Safra Catz. Safra Catz makes a big splash in the back half of the book. He hires her in the late 1990s. She basically becomes—I think it’s officially his chief of staff—but she becomes his right-hand woman.

A lot of the people who were his number two get pushed out by him. But the end of the book is the author speculating on who will take over Oracle when Larry leaves, right? And Larry—it’s the early 2000s, so he’s in his early 60s at this point. It’s not crazy to think, “Hey, who’s going to be next?”

The author basically dismisses Catz and says she’s super loyal to her boss, but she doesn’t have what it takes to be the CEO. Then Larry, right on the heels of that, comes and says, “Here’s exactly what I want in the CEO who takes over for me.” And it is completely not Catz’s description. Then he lists 2 internal candidates who have no similarities with Catz’s skill set or anything.

I just found that, with the benefit of hindsight, so interesting. I’m sure people get over it. Catz gets promoted to be the CEO of a Fortune 5 company, so I’m sure she gets over it. But I thought it was really interesting to see her getting dismissed as the CEO when, with the benefit of hindsight, we know she kind of wins the totem pole. So I wanted to end on that because I found it hilarious and really interesting with the benefit of hindsight.

Yeah, I think it’s just really hard to tell how people will evolve and how companies will evolve. I think you could look at Apple’s succession as a really interesting example of this, where if you had asked who could fill Steve Jobs’s shoes, you’d only be thinking about people who were on the design side or technologists. You would not be thinking, “Well, when you think of Steve Jobs, you think of continuous improvement in COGS and reduction in working-capital needs by effectively outsourcing to East Asia.”

On the other hand, that was more of what Apple needed at that time. I think sometimes you want companies where you’re not trying to have every CEO fit the same mold, because if the previous CEO is really good at solving whatever problems the company faces, the problems that the company cannot solve will be the things that that CEO is bad at. That selection effect means that you probably always want to have successive, differently good CEOs.

There may be some fun internal-politics story. It would be very interesting to read a memoir by Safra Catz’s own chief of staff, or someone who’s been able to follow her calendar and know how she operates. If you write a book like this, you’re interviewing a bunch of people, and the people who have the time to talk to you for hours are often the people who got fired. So, of course, they have an opinion on what they did well and on how unethical the people who are still at the company are.

There was this bit about how, for a while, basically all communication was going through Catz, right? So she kind of knows where the bodies are buried and does have some internal leverage. If the head of EMEA sales is constantly telling her, “Things are not really working this quarter. It looks really bad,” and she doesn’t tell Larry, then Larry finds out that sales in Europe were actually way below expectations. That may make that person look bad, not her.

I have no idea. She actually seems like a really effective manager, just a completely different kind of manager than I would have expected. But she was also in a position to win a lot of those internal political fights. So, yeah, very ambiguous. My baseline belief is that she’s just a very organized operator, and that at Oracle’s scale today, you just don’t need that many big, bold dreamers. You probably need someone who can balance frequent buybacks and growth investments.

No, look, you hit the nail on the head with her and the internal struggles, because even Larry’s number two at the time says, “Hey, I think he’s getting ready to leave Oracle.” He basically says, “I started telling Larry, ‘I need one-on-one time with you.’” And Larry was just putting everything through Catz.

I do think it’s interesting, on your point about the weakness of the person, that Larry very much sticks around. He’s still the chairman of Oracle. He’s still the CTO of Oracle. So it is interesting that he’s still looming large, and maybe that helps her work with him.

The other interesting thing is, if I remember correctly, she’s named co-president with Mark Hurd and, I think, co-CEO at the time. Then Mark Hurd passed away in 2019, and Catz becomes the sole CEO. Mark Hurd is just interesting. He was at HP and was fired from HP amid sexual-harassment claims involving a subordinate. It’s like, hey, that doesn’t fly at HP, but maybe at Oracle, back to what I was saying, maybe over at Oracle it kind of does fly.

Any last thoughts on the book? Look, I don’t know whether listeners would ask, “Hey, would Andrew recommend reading this book?” I don’t want to speak for you, but I’d be like, “I don’t know, man. You have to skip the boating parts.” But it was really interesting. There were a lot of things that, as just somebody who’s puttering around drooling out of his mouth all the time, I would never have thought about internally, and that I understand a lot more about now.

No, that’s exactly how I would read it. The book has this kind of weird, almost novelistic presentation, where the actual plot of the book is basically 6 or so quarters in the immediate aftermath of the dot-com bust. As many novels do, you tell this tightly constrained story: here’s where all the action happens. Then periodically there will be a flashback, someone tells a story, or we get some context, and so you weave it into this full story.

I was kind of expecting Larry Ellison—the first sentence of the first paragraph would be, “Larry Ellison was born here, and he grew up in this city, and he went to grade school, and his friends remember him being rambunctious,” or whatever. But no, it’s like—I think the very beginning, other than Larry negotiating with the author over what the book will be, also has this weird postmodern quality. It’s kind of like Gonzo journalism, because this guy is often in the room, and he’s talking about being on the plane with Larry, being in the car with Larry, and so on.

You just open with Larry getting in a car and driving off to talk to another set of customers.

The book opens with HealthSouth, right? Yeah.

Yeah. No, I can’t remember if it was last month or 2 months ago, but we did The Snowball, and I was interested in the difference. The Snowball starts with Buffett’s current career in the first chapter, but then it quickly goes, like you said, to Buffett being born.

It was interesting that The Snowball really hits on Buffett’s relationship with his mother and father: a terrible relationship with his mother, and a father whom Buffett idolizes, right? With this book, it doesn’t mention Larry’s beginnings at all until really the middle. It’s kind of crazy because Larry is adopted, and his adopted father is basically—not that he doesn’t want a relationship with him, but he clearly doesn’t care for Larry.

In the hands of another author, you start with that and have Larry as the guy with a chip on his shoulder throughout the entire book.

The boating is more important than that. I just thought it was an interesting choice, and it also puts a little bit more of it in perspective, I’d say.

I could imagine Larry having talked to the author about this and basically saying, “My childhood wasn’t great. As soon as I could, I moved to California and built the life I wanted to have.” So at that point, the childhood is just this background context that he’s not really thinking about.

I think that does make sense as a reason that we wouldn’t have seen a ton of this detail. I always go back and forth on whether—I’m always impressed when a business biography is able to track down someone who went to middle school with this person and is able to talk to them.

That makes me hope I never get a biography written of me, but—

I think I’m safe. But then I also asked myself: I’m sure there were a lot of people who were similarly talented or driven, and they just got unlucky. Maybe the year that they would have started their amazing database company, a family member got sick, and they had to spend years caring for this person. That’s why there’s not a book about them, even though they’re every bit as cool and interesting.

On the other hand, it’s interesting to get that context because part of what you get from this is that this person was unusual from very early on, but in a more understandable way. It wasn’t like they were a complete alien who was just beamed down to this planet. They’re on the same spectrum as the rest of us in a lot of respects.

When I read this book, Larry Ellison comes off well in it, and I do think there is—you mentioned gonzo journalism—I understand Larry didn’t get to reject anything in the book, but I think the journalist has a real soft spot for Larry. But when I read the Buffett biography, one of the quotes I said to you was, “Dude, he was 12, and people were calling him up and saying, ‘I have a warehouse full of goods by you. I need you to move this stuff,’” and trusting Buffett. I was like, there was something clearly unique, almost immortal, about him.

There was nothing about Larry Ellison like that. He’s very charismatic, but a lot of it is kind of right place, right time. And, again, it’s funny: he says Microsoft was willing to break the rules. From the founding of Oracle, where you’re working at a company and you direct a contract to yourself, he’s very slippery with rules, I would say.

There was just nothing in here that I would say, “This man is a god among men.” I would be surprised if anyone would read the Buffett book and not think, “Hey, there was something really unique about this guy from a very early age.”

Yeah. Buffett definitely has some of those really outlier traits where it’s just like reading about early Mozart, where he sits down and plays and you’re like, “How long have you been playing the piano?” He’s like, “Oh, that’s what this thing is called.” It’s something like that. And, yeah, you do have a different vibe with Ellison. He’s, like I said, really charismatic.

Clearly, this is something I had underestimated about him before rereading the book: he actually has this definite vision of the future. Whereas I always thought of Oracle’s vision—the future that Oracle envisions—as the quarter they’re guiding for, and that they don’t actually have some broad technological vision. And yet you read it, and he basically correctly predicts how we will use computers today. He has to shove this narrative down people’s throats.

Maybe this is too much of a tangent for this late, but he has this line early on where he tells everyone, “You should not customize your software for your business process. You should actually customize your business processes around my software,” which is an incredibly arrogant thing to do. Except we clearly do that. The existence of the Like button has just rewired everybody’s brains, and we’re used to thinking in terms of status updates.

There are so many companies where their name has become a verb, and it just changes the scope of what you can do. It changes how you think and behave. So he actually got that part right, too: the software is actually going to tell you what to do, and it’s going to tell you how to think. You’re still thinking, but you are riding on tracks that are written in code. That was one surprise from the book.

In terms of whether it’s useful and who should read it, I think anyone who’s dealing with large IT projects or just wonders, “Where does this money go?” Like, you read quotes about how many people in IT are employed at large banks, and you ask yourself, “What do they do all day?” There are only so many ways for Chase to check whether my balance updated correctly if I bought something at Starbucks.

It is clear that these are incredibly complicated systems, not just complicated technological systems, but complicated human systems. And maybe if there is something extraordinary about Ellison—some factor in which he’s just a massive outlier compared to the rest of us—it might be his fluency in translating between those layers.

Sometimes that is just a technological limitation of the relational database model: it means that your company has to do something different, and this person needs to get fired. If you can actually make that translation and then have the willpower to implement it, you can do pretty extraordinary stuff.

Especially if you happen to be born at a time when you are young enough to found a company when Moore’s law is really kicking into high gear, and then your company is well positioned for a world of networked computers, where there are centralized data centers, not one really big hunk of iron from Sun Microsystems or IBM.

He did—I think some of this was probably luck. I don’t think he started Oracle thinking, “Well, when every computer is networked, databases will be more important.” But he certainly seems to have seen that happening sooner than other people did and made the right calls based on that. So, pretty impressive. It’s just crazy that he’s still kicking. It’s so great.

Look, I think it’s chapter 19: all he’s focused on is longevity, so maybe that makes sense. We’re going to have to end it there because I think you’ve got 3 kids all on your own. I don’t know how you’re doing it, but I’ve got to run as well. But this has been great.

I was worried when I was reading it. And then we talked for an hour, and I have 5 different notes that you were just saying that I wanted to follow up on, but we’re going to have to put it all there.

Next month, I will tell you: if we had recorded this podcast one week ago—we’re recording on July 25th—if we had recorded it July 18th, things were getting toppy enough that I was going to be like, “Byrne, it’s time. We have to do a bubble book. We have to bring out Boom by Byrne Hobart and discuss how manic things are getting.”

I think things have kind of cooled off just a touch since then, but I’m telling you, man, we’re on the verge. We’re close to pure mania, and if it tips just a little further, we’re busting out Boom for next week. To all our listeners, if you have ideas for next month's book club, we'd love to hear them. Byrne Hobart from The Diff and Capital Gains. This has been great, and I’m looking forward to chatting next month.

Absolutely. This is great. A quick disclaimer, nothing on this podcast should be considered investment advice. Guests or the hosts may have positions in any of the stocks mentioned during this podcast. Please do your own work and consult a financial adviser. Thanks.

Absolutely. This is great.