Legendary Investor Bill Gurley on Investing Rules, Insights from Jeff Bezos, Must-Read Books, & More

Valuation multiples are always a crude proxy and dangerous to use; if multiples must be used, 10X should be considered amazing and an upper limit, and anything over that is silly.

“Valuation multiples are always a hack proxy. Dangerous to use. If you insist, 10X should be considered AMAZING and an upper limit. Over that, silly.”

Twilio's valuation went from 70X gross margin to 3X gross margin in a very short window—a radical reset demonstrating how valuation multiples can completely collapse when sentiment shifts.

“Still, Twilio went from 70 times gross margin to three in a very short window. I mean, talk about valuation reset, that is just radical.”

Writing as a thinking tool: when you write a six-page paper (as Bezos requires for Amazon meetings), you are forced to think more deeply than a five-slide PowerPoint because it's harder to hide weak reasoning in prose; writing clarifies thinking by forcing you to articulate premises and defend them.

“By the way, mirror back to what we talked about earlier about writing and thought process, if you're forced to write a six-page paper, it's much harder to put that together than it is a five-page PowerPoint. It's easier to leave stuff out. You really have to think through everything.”

Gurley believes that if Meta/Facebook shut down the VR effort (currently spending 5-10 billion per year), not only would profitability soar, but the stock would double; Wall Street agrees with this assessment, suggesting the VR investment is value-destructive rather than value-creative.

“If they shut down the VR effort, not only– Well, the profitability would soar because they're spending real money, like five to 10 billion a year, but I think the stock doubles. Doubles.”

Investor bedrock comes from reading widely (Buffett letters, Peter Lynch, Mike Mauboussin, Burton Malkiel) and building an understanding of financial history that constrains overly optimistic or pessimistic biases.

“I think any investor starts with just building bedrock and that comes from reading. And there's a ton of books in history. You can go through all the Buffett letters as an example. You can read Peter Lynch's One Up On Wall Street. There's my good friend Mike Mauboussin who has put out some amazing books that are a lot more nuanced about stock prices, that kind of thing. A Random Walk Down Wall Street by Burton Malkiel.”

Howard Marks and Stan Druckenmiller are two of the very few individuals known for being successful at macro investing; most experts including Buffett say macro is impossible and shouldn't be attempted, but these two have demonstrated sustained success in macro through different approaches—Marks through long tenure in bond markets, Druckenmiller through single-individual macro bets and recent success predicting inflation.

“There aren't many people, if you study financial history, most people in Buffett included will tell you, 'Macro is impossible. You shouldn't even try.' And the two individuals you mentioned are two of the only ones that are known for being successful in macro investing. Howard mostly by being one of the most successful and longest tenured investor in the bond market and Stan for taking more kind of single individual bets that are macro in nature going back to his success with Soros, he's become re-famous in the past 12 months for predicting the inflation situation we're in and being very loud about it.”

Facebook (Meta) in spring 2022 was trading at 14X GAAP earnings while growing at 23 percent; this was a historically low multiple for such a successful company, contrasting sharply with Coca-Cola which traded at 30-35X earnings but only grew 3-5 percent, suggesting Facebook was undervalued on earnings-based metrics.

“Facebook trades at 14X GAAP EPS… and is growing 23%. What earnings multiple are you assuming?... here's one of the most successful companies of all time that is producing massive amounts of positive cash flow and GAAP-audited earnings that's trading at a very low multiple of those GAAP earnings... they're trading at 14. Or you may not even have earnings... Coke is less than 10 percent growth, like five or three.”

Institutional policies that were developed on university campuses around individual accommodation and anti-fragility have crossed over into corporate environments, where they are particularly inappropriate because they undermine the performance and coordination necessary to run a high-functioning company.

“This happened in the university and this did get out into the companies and so these companies were being basically held accountable for delivering an experience for the individual. It's like your goal as a company is to help the individual have a great life. And I think it's very hard to be high performant. Imagine if Coach K. had that problem, like he had to make everyone make sure everyone on the team felt happy and safe and included and that was true north. It would be very hard for him to be performing.”

Earnings quality relates to cash flow; a company might have strong GAAP earnings but weak cash flows due to timing differences or other factors, creating a divergence between reported earnings and actual economic performance.

“Earning qualities typically relates to cash flow. So you might have really good GAAP earnings, but because of different factors in your business, your cash flows may not be nearly as good. It could be timing differences, those kind of things.”

Coach K. (Mike Krzyzewski) at Duke built a culture around three tenants: expect the very best from each player, team first (if you need to be an individual, don't be here), and a singular goal (winning the national championship); this type of high-performance culture has become harder to implement in 2020-2021 because companies are held accountable for delivering an experience for the individual rather than focusing on collective performance goals.

“Coach K. told us that he expected the very must out of us. Each individual had to perform at their highest level. He said, 'It's always team first. If you need to be an individual, you don't need to be here.' And Shane was talking about how people remember the winners more than they remember whether you were the third or fourth scorer on a team. They remember the winners, and I think that's true in startup world as well. And then three, he said, there's a singular goal for this organization and it's to win the national championship...if an internet entrepreneur stood up and said those things out loud about what he wanted from his company, he might get canceled. Company first, you have to perform at your absolute best and the only goal here is the one goal of the corporation, we're all on a team together. We wandered to a place that's very different than that.”

Brian Armstrong at Coinbase and Tobi at Shopify have publicly taken stances that the company's one objective is the company itself, and if employees are hyper-passionate about something else they need to discuss, they should go do that instead, reestablishing performance accountability that had been lost.

“Brian Armstrong at Coinbase has very publicly spoken. I think Tobi at Shopify took a little more nuanced stance, but the same stance that like we have one objective here, which is this company. And if you're hyper passionate about something else that you got to talk about it all the time with this company, maybe you should go do that. But this performance awareness had kind of gone away because of everything I talked about.”

A CFO hired for OpenTable quit because his financial model showed the business would never work, having frozen market penetration at 17 percent based on his experience in retail where no business gets more than 17 percent market share; Gurley believed in network effects and predicted 99 percent penetration instead, and was vindicated when OpenTable later filed its S-1 with massive cash flow.

“So this was in '99, so this is a 23-year-old story, but I had successfully recruited a CFO from a public company, which you could do back in those glory days to come into OpenTable. And one day I showed up early for a board meeting, and the CFO comes to me and he says, 'Bill, I'm going to quit.' And I said, 'Okay.' I go, 'Why are you going to quit?' And he goes, 'This business will never work.' And I said, 'Okay, why will it never work?' And he says, 'Well, my model says it'll never work.' So I said, 'Show me your model.' So we look at the model and I dive in, and he has frozen penetration in each city at 17 percent... And I go, 'Why'd you freeze it at 17 percent?' And he said, 'Oh, no one gets more than 17 percent market share, all the businesses I've worked with.' And because I believed in network effects, I was like, 'We're going to get 99. We're not stopping at 17, we're going to get 99 percent.'”

Michael Mauboussin, a food analyst at Gurley's firm, was a natural learner who shared the Return on Invested Capital (ROIC) framework with analysts; Gurley applied this framework to all his covered companies and found that Dell had unusually high ROIC numbers (20 to 1 compared to the rest of the industry) despite the stock being in poor condition, which contributed to Gurley's strong buy recommendation and 100X return in the public markets.

“He had just read a bunch of books on a framework called Return On Invested Capital, which Stern Stewart had published on... And he was spreading it. It was the word proselytizing through the analyst group. And I was a sponge at that point in time. So I took the framework and ran it on all of my companies. It turned out just by happenstance that Dell stood out like a sore thumb with ridiculously high ROIC numbers versus the rest of the industry. Like night and day, like 20 to one, wasn't even close.”

Benchmark focuses on several repeatable venture models: network effects, open source, social networks, and SaaS (software-as-a-service), among others; they have been successful with multiple companies in each category and have developed intuition about what works, though they remain cautious because rules can mislead.

“There's like four or five different ones that we look for, network effects, open source, and you get good at it. You start to understand what works, what doesn't, what leads to success, what doesn't. I mean, there's nuances. In open source, with that business model, which is you're basically packaging support and reliability, because someone could just download it for free so how do you charge?”

Gurley became an All-America Research Team analyst for Institutional Investor largely because three top analysts in his category retired within one year, creating an opening; he then used a framework of contacting salespeople to ask for 20 introductions to clients who would spend 30-45 minutes telling him what they wanted before he had any research output, which allowed him to understand what clients valued.

“But when I showed up on Wall Street in one of the very first weekly meetings, they introduced us to the sales force... And somewhere in my youthful wisdom, I decided to ask each salesperson, is there one client that will spend 30 to 45 minutes with me as a new analyst and just tell me what they want. I'm just going to ask them questions. I'm not going to have anything for them. I just want to know how I can serve them best. And I did 20 of those interviews before I had started the job. And that's a roundabout answer to how do you get on the list because I knew what they were looking for at that point.”

At OpenTable, network effects were working when one salesperson in San Francisco—where they had 90 percent market penetration—closed 35 restaurants in a month compared to the target of four per salesperson, demonstrating that the last 10 percent of restaurants in a saturated market were easy to close because consumers were already on the platform.

“And so I asked the question, 'Who is that salesperson?' And OpenTable started in San Francisco, and we played local game. We didn't go everywhere at once. We built liquidity city by city. Anyway, at the time, that salesperson was the one salesperson left in San Francisco where we had 90 percent penetration. And so the 35 were coming out of that last 10. And that individual was basically taking orders. And that to me was like, 'Oh, yeah, the network effects really working here.'”

Jeff Bezos is probably the best entrepreneur Gurley has ever been around or gotten to know, and it's remarkable and multifaceted; his ability to build organizational frameworks that embody his beliefs at radical scale is underappreciated.

“I mean he's probably the best entrepreneur that I've ever been around or got to know. It's remarkable and it's multifaceted. Here's one that I think is not well discussed so he has a bunch of traits that make him a great entrepreneur. The company today is at such a radical scale that there's no way, and I mean he's in the chairman role, he's not touching all the decisions, he's not touching all the product decisions, he has built a organizational framework to take what Jeff Bezos believes and run the whole company that way and that's not well dissected, not well understood.”

Bruce Dunlevie at Benchmark learned to 'have a really big tent' and be super open-minded about information sources because it increases the chance of discovering opportunities; he developed deep relationships within companies and when an engineer from a portfolio company approached him with eBay, he made an intro to Bob Kagle who was more excited, demonstrating how equitable partnership allows deal flow to follow the most passionate partner.

“From Bruce Dunlevie, is just have a really big tent. We're in a networking business where you're trying to look under every rock for the next possible deal. And cutting off avenues of information or flow is just a really stupid idea. And so Bruce had been involved with a company that I think was in a software tool space, the name escapes me. We could look it up. And had developed a lot of friendships, went down the organizational chart at the company. And one day, one of the engineers from that company comes to him and says, 'I'm working on this marketplace thing called eBay.' And he was from a different industry, Bruce, but because he had developed that relationship that came in, now, it turned out that Bob Kagle was much more excited about eBay than Bruce was. And so Bruce made that intro because we're equal partnership, that's all cool.”

Gurley calls the use of open standards by large tech companies 'defensive corporate strategy' and sees it as an emerging pattern where incumbents use open source not altruistically but to prevent competitors from monopolizing markets and to reduce supplier leverage.

“I call it defensive corporate strategy.”

Benchmark's founding partners created an equal partnership structure (unlike most business partnerships which are hierarchical), where each partner has equal voice, equal ownership stake, and equal upside incentive; this structure creates emergent properties where partners root for each other's success rather than compete, which enables contrarian thinking and risk-taking in partner meetings.

“The founding partners of Benchmark did something that hadn't been done before in venture, which is they created an equal partnership. And most business partnerships, and this includes private equity and law firms and real estate firms, there's a hierarchy. And the people that have been there the longest sit at the very top, and they take an outsized amount. And some of our founding partners had worked at those firms and felt like the young people did more of the work, and therefore it wasn't conducive to the right type of internal behavior. And so they created this notion of an equal partnership.”

When Benchmark invested in OpenTable with only three restaurants, they had to believe that if they acquired enough restaurants, consumers would follow, and if consumers came, restaurants would have to participate; they had to install broadband alongside the PCs sold to restaurant owners, which would normally be a "don't do that" decision, but network effects fundamentals made the asymmetric bet worthwhile.

“When I met with Chuck Templeton, the founder of OpenTable, and he had three restaurants. You had to believe a lot to get from that point to the global phenomenon that it became. And the bet that we talked about making when we said, 'Okay, let's go do this thing,' is 'If we get enough restaurants on this thing, then the consumers will come. And if the consumers come, then people will have to get on.' And that's what happened. But you had to believe it because otherwise, at the time, we were selling PCs to restaurant owners, and guess what? They didn't have connectivity at the time. So we had to partner with someone to get broadband installed, which wasn't easy.”

Most venture capitalists don't have a liquidity event until year eight or nine, so it's easy to doubt yourself; three years into a portfolio, companies resemble "12-year-olds becoming 13-year-olds, your whole portfolio's like that with acne all over their face"; partners can become anxious and lose confidence, so support from more senior partners saying "it's going to be okay, get back out there" is more helpful than people realize.

“Unless you're super lucky... they don't have a liquidity event till year eight or nine. And so it's easy to doubt yourself. And three years in, people like to use the child age analogy, you've got 12-year-olds becoming 13-year-olds, and your whole portfolio's like that with acne all over their face, and you can really lose confidence. So one thing that a lot of my partners did was just like, 'It's going to be okay, get back out there. You're good. You're doing fine.' That kind of thing, which was way more helpful than you could possibly imagine because the anxiety was spiking.”

A professor at Santa Fe gave a presentation on electric grid problems in North America and concluded that 'the best solution is smaller communities that are loosely coupled'; Gurley realized this same framework explains why the Euro (tightly coupled) is problematic, and why distributed systems that aren't hyper-integrated avoid global failure modes while remaining scalable.

“The professor that presented it at the end said, 'The best solution is smaller communities that are loosely coupled.' And that is a really interesting — you could use that framework. I walked up to her afterwards and I said, 'You just explained why the Euro's a bad idea, because you tightly coupled this thing too much.' And you can just imagine even a computer system that's distributed, that learning could have other applications. If it's too distributed there's no scale. But if it's hyper integrated, then you have this failure problem, this global failure.”

If Gurley were to give a lecture on cultivating anti-tribalism, he would start by demonstrating that people are turning off their brains; then explain why people want to be intellectually consistent and what intellectual inconsistency means (e.g., excusing spouse infidelity when your side does it but condemning it when the other side does).

“I think the first thing I would try and do is just highlight the fact that people are turning off their brains. The way that people have proven how confirmation bias works, how loss aversion works. There's 20 different cognitive biases that we're all aware of. The first thing I would do is just run some — there's been stuff done. Then, why does anyone want to be intellectually inconsistent? If you think it's an ends to a means, then you're just in a fight. You're just in a fight. I don't even want to have a discussion with someone if cheating on your spouse and philandering is okay when your side does it, but it's horrible when the other side doesn't like, right. What's your principle?”

Benchmark changed their decision-making framework after missing Google by adopting the principle 'What could go right?' drawn from Matt Ridley's 'The Rational Optimist,' which better aligned with the asymmetric payoff structure of venture investing.

“And around that time I remember we used to give a book out to our LPs, our limited partners, our investors at every annual meeting. And I think Bruce had just read The Rational Optimist, which is a Matt Ridley book. And he started using a phrase at our partner meeting, 'What could go right?' And because of this asymmetric outcome thing where you could make 10,000 times your money and only lose once on the downside, it was the right frame of mind.”

Jeff Bezos tweeted positive comments in response to Gurley's spring 2022 valuation reset thread on Twitter, providing external validation of his analysis that the industry needed to reset its mental models.

“Jeff Bezos, I think either retweeted or replied to that tweet thread and with was very complimentary.”

In spring 2022, Gurley tweeted that an entire generation of entrepreneurs and tech investors built their perspectives on valuation during a 13-year bull market, and the 'unlearning' process would be painful, surprising, and unsettling; he anticipated denial about what was happening.

“'An entire generation of entrepreneurs and tech investors built their entire perspectives on valuation during the second half of a 13-year amazing bull market run. The 'unlearning' process could be painful, surprising, and unsettling to many. I anticipate denial.'”

Bezos was asked 'When do you stop an experiment?' and answered 'When the last person with good judgment gives up'; this is not how other big companies operate, and contrasts with typical big company experiments that are killed after one test failure.

“Someone asked him, 'When do you stop and experiment?' And he said, 'When the last person with judgment gives up.'... This happened in the university and this did get out into the companies and so these companies were being basically held accountable for delivering an experience for the individual.”

Shopify's Shop app was very cool because Shopify is a B2B platform company but created a consumer network effect product; very few companies can successfully cross over from B2B to consumer with network effects, making this a remarkable achievement.

“I thought what they did with the shop app was super cool. So the previous Christmas I started noticing websites I'd never been to before, knew who I was and allowed one click checkout. And very few companies can make their business as a B2B company and then have this crossover product, which has a consumer network effect. And they did that with that app and I thought it was just super cool that they pulled that off.”

Amazon, while a company worth tens/hundreds of billions and running mature businesses, ran an early same-day delivery experiment using Uber, burning phone numbers and manifests to drivers—something most large companies wouldn't do because of accounting/governance concerns; this demonstrates Bezos's institutional commitment to experimentation and risk-seeking.

“This is a company that's worked tens, hundred billions of dollars that is running an experiment on top of Uber and I know for a fact that most of the companies I've worked with that have gotten over 20 or 30 million in revenue would not run that experiment because someone would say, oh, we won't know how to do the accounting, that's too much of a hack, like whatever. But this large company was super comfortable running this kind of hack experiment on this other company.”

Tobi Lutke (Shopify CEO) told Gurley a brilliant framework: when dealing with a problem in a meeting, he starts by saying 'The one option we know we're not going to leave doing is the status quo'; this removes the default and forces real decision-making rather than inertia.

“He said, 'Whenever we're dealing with a problem and we call a meeting to talk about the problem, I always start with this structure. We are here to solve a problem. So the one option that we know we're not going to leave the room doing is the status quo. That is off the table. So whenever we finish this meeting, I want to talk about what option we're taking, but it's not going to be what we're currently doing.'”

Autonomous vehicles should be open source because safety improves, communication layers are better, and test suites can be shared; treating traffic light detection as an ML problem is 'really stupid' when it's a state machine (red, yellow, or green) that could be communicated directly; academia and corporations could work on the same standard, making government involvement easier.

“First of all, I think autonomous vehicles should definitely be open source. I think everyone benefits, safety's higher, communication layers are better. The idea that you would use artificial intelligence to figure out whether a light is red, yellow, or green is really stupid because it's a state machine. It is in one of those three states and that could be communicated into the software. You don't need to infer that, that's a known thing, but you need a common language. And your test suites could all be used by everyone. Academia could be working on the same thing that the corporations are, safety's higher, easier for government to get involved if it's a single standard.”

Global nuclear energy standards could reduce costs and improve safety; if the globe had a standard for fission-based nuclear reactors, we would get lower price points, safer products, and more participant involvement, contrasting with current situation where high costs are due to regulation, not the actual product.

“I look at the state of nuclear energy where the costs are high because of regulation, not because of the actual product and I wonder, what if the globe had a standard for fission-based nuclear reactor? Wouldn't we get to lower price points, wouldn't we get to safer products, wouldn't they get more involved?”

Benchmark operates under a constraint where board seats are the limiting factor more than capital itself because they choose a strategy where they place their money with their work and involvement, becoming the largest shareholder and serving on boards, so they cannot do 20 investments 'like not well'.

“And we have another challenge, Tim, which is we've for a variety of reasons have chosen a strategy where we don't let our money walk around without our work product and our involvement. And so we go on a board if we make an investment and we usually become the largest shareholder on the board. And as a result, our limitation is our board seats more than the capital actually.”

HackerOne was an obvious 'yes' investment because when it was founded, only four companies (Microsoft, Google, Facebook, Mozilla) ran white-hat hacker bug bounty programs; the founding team came from Facebook's program and the insight was tautological: if it works for these four companies, it obviously works for every other company in the world, so Benchmark said 'yes' immediately.

“At the time of HackerOne's founding, there were four companies that used white hat hackers to make their sites more secure. Microsoft, Google, Facebook, and Mozilla, they were the only four. No one else did it... And when they presented, we didn't even discuss the company. The minute they left, we were like, 'Okay, how are we going to close this?' Everyone had just jumped to 'Yes,' because it seemed so tautological, that there's no way this thing's great for these four companies and no one else.”

Twitter's mechanisms incentivize polarization through its feed structure, reward system, and engagement model, yet Gurley posts on Twitter because it offers unparalleled ability to get close to experts and leaders in fields through direct conversation and DM relationships.

“I find Twitter just to be super fascinating as your ability to get close to experts, leaders in your field, like — It's unbelievable. More so than LinkedIn. And there's a chance that some reply you give to them, they might like or they might follow you. And now you, you've developed a mentor or a peer partner. It's really just shocking, amazing the amount of information you can take out of this thing if you use it properly.”

Gurley intentionally expanded the distribution of his weekly analyst pieces as far as he could, leveraging the sales force for fax numbers and developing industry relationships through investor days and company meetings, then found a major distribution hack by attending Stewart Alsop's Agenda conference where he bought a Palm Pilot loaded with contact information for 400-500 of the most influential people in tech and spammed his newsletter to them at approximately 70 cents per name.

“I leveraged the sales force and get them to give me fax numbers. That's that's where I started. I then started developing industry relationships, which is important because you're covering these companies, you start going to investor days. The buy-side is talking to these companies as well. And so I started getting some of them on board. And then probably the most successful hack of my career happened when I was invited to attend Stewart Alsop's Agenda conference... And I ran some quick math. People weren't really doing cost of customer acquisition back then, but I think it was like 70 cents a name or something like that of the most influential people in the entire tech industry. And so I bought the Palm Pilot, I took it home, and I spammed the 400 or 500 most important people in the tech industry with my weekly newsletter.”

LinkedIn has stalled for about ten years; there's potential for a new take on LinkedIn using page rank for people where private opining about who's smartest on particular topics (rather than public skill claims) could develop unique product experiences based on who each individual trusts.

“I'm highly interested in people, feel free to reach out in any variation or new take on LinkedIn. I just think it kind of stopped and it stopped 10 years ago... What if you had people maybe privately opining on who they think is smartest on particular topics? You can develop a really cool product that would be unique to each individual. Because what you see, Tim, would be different than what I see because it starts by who you trust and then what you see is there”

Twitter could build a top-down version of itself (rather than feed-based) that uses algorithmic scoring to identify experts on particular topics (e.g., a page for each stock showing top stories from Twitter about that stock, pulled by identifying who the 'axes' are on that topic), and could extend this to other domains like sports teams, presenting curated expert content rather than algorithmic feed noise.

“What if you had a page where, for each stock, you had a list of the top stories of the day, and that could be pulled out of the Twitter feed by knowing which people are the axes on that individual stock, which you could infer simply with the data that's already in there. And so you could imagine that for sports teams. So I could have Twitter sports that is a top-down version that's using the information in the feed, but then presents it more like a standard newspaper would, if you understand what I'm getting at.”

Aswath Damodaran made a classic TAM analysis error by calculating Uber's ceiling as the size of the existing taxi market, but Uber created a product 10X better than taxis (convenience and availability), which expanded the TAM by 20X in San Francisco alone, illustrating that TAM conservatism causes bigger losses than TAM optimism in venture investing.

“He basically took the taxi market and said, 'That's the upper limit,' and that's just the wrong, like, math we made, or Travis and the team made this thing so convenient and so available that it was a product that's 10x better than the taxi market. And by the time I wrote that, I already knew that Uber in San Francisco was 20x bigger than the taxi market in San Francisco. So I already knew he was wrong with that analysis...I have found you get into more trouble with this kind of TAM conservatism, then it hurts you more than it helps you as an investor.”

Complexity by Mitchell Waldrop, about the rise of the Santa Fe Institute, is the book Gurley has gifted most frequently; it introduces the concept of complex adaptive systems and helped Gurley understand network effects.

“The first one that I've gifted the most is called Complexity by Mitchell Waldrop, which is about the rise of the Santa Fe Institute... That book introduced me to network effects.”

Mr China by Joe Studwell is a fantastic book about someone who went to China in the mid-1990s and started a fund to privatize industries; he was from London, got his head handed to him, but survived to write a humorous story about it; it's valuable for understanding business risks in foreign lands.

“Mr China is fantastic. He went to China in the mid '90s and started a fund to privatize a bunch of industries. He's from London and got his head handed to him and survived enough to write a humorous story about it.”

Startup by Kaplan is a fantastic book that chronicles the early portable computer industry where every single company failed despite having the best venture capitalists, best advisors, and best executives; Kaplan maintained a diary on cassette tape on his commute, which provided exceptional detail; it's valuable to read books of failure alongside books of success.

“There's a book I love called Startup by Kaplan... He was in the — what do they call that world? Because it's where General Magic was. It was like the first portable computers. Every one of them failed, every one, but they all raised massive amounts of money. He, on his ride home every day, had a cassette tape and left an archive or a diary, which makes the book so good because the detail is fantastic. But they had the best venture capitalists, they had the best advisors, they had the best executives, and the executives that were in — this was GO Corp, I think, that were there. All went on to do amazing things. But this company failed hard, failed hard. And I think it's nice to combine a book of failure with all the books of success.”

Gurley is working on a book about chasing your dream job and how to succeed at it; his research shows 70 percent of people have career regret, which is a huge number indicating widespread dissatisfaction; the book builds on studying biographies of unusual people (Bobby Knight, Bob Dylan, Danny Meyer) to identify similar patterns.

“There's a speech I gave at the University of Texas that you could put in the show notes... And I've done some research since I gave the speech because people have encouraged me to turn it into a book. And some of the polling we've done show 70 percent of people have career regret. 70 percent, which is a huge number. And so we're doing some more work to better understand that and how people end up in that place. But the punchline, which I hinted at earlier is I just don't think there's ever been a better time to have a self-determined job process if you want.”

There is a gap in the market for an 'interest graph' company—a website linking everyone tied to specific interests; companies like Pinterest, Quora, Twitter, and Reddit come close but none have really nailed it; if created successfully, such a platform would provide unique unlocking for people by connecting them with others at their skill level and would have exceptional advertising performance due to precise audience bucketing.

“Everybody talks about something called the interest graph, and they wonder why there isn't an internet website that kind of links everyone that is tied to a specific interest. And there are companies that people talk about as being close, like Pinterest or Quora or Twitter, but I don't think any of them have really pulled it off... And if you did, you would have this combination of really cool unlock for people because you'd get connect– if you were into quilting, you'd be immediately connected with everyone else that's on your level and you could imagine that kind of thing. But then the advertising performance would just be off the charts because you've kind of bucketed everyone into these places.”

Zuckerberg's sunk cost bias and confirmation bias make it unlikely he will cancel the VR efforts despite years of people telling him the premise is wrong; historical examples like AWS and Android showed big bets that started off-mission but succeeded, but the VR effort has already exceeded the development costs of either and lacks comparable unit economics.

“It's funny, a lot of people love to discuss this, and what's that other– There's a bias when you get pot committed, to me that's confirmation bias too, but if you've already bought something, you like it way more...he's had people telling him what I just said for two years now...I think the two most amazing are AWS and Android...you've already run the clock, you've already spent way more than either of those did...And you just don't have the numbers.”

Meta's AI tools and WhatsApp are more promising than VR investment; WhatsApp has a similar role in India that WeChat has in China, and Meta has accomplished remarkable things with it.

“There's really good buzz on some of their AI tools and the world's super excited about that... WhatsApp has some interesting things going on... if you read about WhatsApp in India it has a similar place in the world that Tencent with WeChat does in China, they've accomplished more.”