Andreessen Backs Musk Management Model to Defeat Corporate Bureaucracy

Andreessen支持马斯克管理模式以打破企业官僚体系 互联网宛如针对体制的X光机而非虚假信息之源，它通过直接向公众揭示系统性失败，正推动一场长期的政治重组。 在这场内容广泛的对话中，科技界资深人士Marc Andreessen和Charlie Songhurst与Stripe的John Collison一同剖析了硅谷的周期性本质，追溯了其从互联网泡沫破裂到现代AI革命的演进历程。他们探讨了西海岸独特的高信任度拓荒文化，将其与波士顿等历史对照样本进行对比，并深入剖析了Elon Musk等标志性领袖的运营策略。此外，他们还讨论了稳定币的兴起、通缩与通胀行业之间的经济鸿沟，以及去中心化媒体如何瓦解中心化机构的权威。 健力士（Guinness）及各啤酒酿造厂代表了社会中存续时间最长的一批机构，其寿命远超常规企业，并保留了深厚的传统。 (0:00) 精准择时宏观市场的顶部或底部几乎是不可能的，即便是最顶尖的对冲基金经理通常也无法准确预测崩盘。 (4:30) 互联网泡沫破裂导致风险投资经历了“失落的七年”（2000—2007年），在此期间创业背负着严重的社会污名。 (8:37) 风险投资受幂律法则主导，需要LP（有限合伙人）保持长期投入，以平滑不同设立年份基金之间的极端波动。 (11:55) 顶级风投的背书如同一笔关键的“信誉过桥贷款”，能让早期初创公司吸引到精英人才和关键资源。 (17:00)

Marc Andreessen and Charlie Songhurst argue that technological progress is governed by power-law dynamics, structural risk-taking, and decentralized disruption, creating sharp divides between deflationary technology sectors and stagnating, highly regulated institutions. They contend that while AI and software multiply human leverage, their rapid expansion disrupts corporate management models and accelerates the erosion of centralized authority. Asymmetric Risk and Power Laws in Venture Capital (11:55) Venture capital operates under extreme power-law distributions rather than public-market averaging, making fund performance dependent on capturing rare outlier successes. Andreessen highlights that losing an investment forfeits only 1x capital (a Category One error), whereas passing on a generational winner (a Category Two error) foregoes thousands of times the investment. He points to Andy Bechtolsheim’s $100,000 pre-incorporation check into Google, which returned roughly 30,000x. Consequently, top-tier venture firms require 20- to 50-year investment horizons and mechanical discipline to keep deploying capital during deep downturns like the 2000–2007 venture stagnation. Silicon Valley’s endurance as an innovation cluster relies on self-selection and high-trust, repeating interactions. Backing from prestigious venture firms provides an early "bridge-loan of credibility," sparking preferential attachment—the Matthew principle—where elite engineering talent, downstream capital, and brand momentum cluster around early-stage companies before commercial viability is proven. Infrastructure Bubbles and the AI Computing Paradigm (40:27) Addressing AI bubble concerns, Andreessen contrasts current spending with the dot-com crash. The late-1990s collapse was primarily a debt-heavy telecom and fiber-optic infrastructure bubble that ran far ahead of consumer broadband, which arrived at scale only after 2005. While today's data center expansion carries comparable debt and capacity risks, Andreessen claims AI provides immediate utility, insulating it from the adoption vacuum that stalled the early internet. Andreessen classifies AI as "Computer Industry V2," representing the first fundamental architectural transition in 80 years from Von Neumann computing to neural networks originally conceived in the 1940s. He forecasts AI adoption will resemble a hardware pyramid rather than a winner-take-all software market: a few frontier models will occupy the top, while the bulk of aggregate computing will occur locally on cheaper, power-efficient, and open-source models deployed across edge devices. Technological Deflation Versus Regulated Institutional Scarcity (66:41) The speakers describe an economy bifurcated into an unregulated deflationary sector and a subsidized inflationary sector. In software, electronics, and light manufacturing, prices continuously drop as productivity accelerates. Conversely, in heavily regulated domains such as healthcare, housing, and education, governments restrict supply while subsidizing consumer demand, driving severe cost inflation. This disparity is compounded by Baumol’s cost disease, where competitive wages from high-productivity tech fields spill into low-productivity, unionized, or civil-service jobs without matching output. Andreessen disputes forecasts of permanent AI-driven unemployment, comparing the transition to the Second Industrial Revolution (1880–1930). While rapid technological extraction caused commodity price collapses and contemporary panic, it ultimately tripled material living standards. He argues AI will act as a hyper-deflationary force multiplier that turns workers into high-leverage specialists and prompts the creation of new, unforeseen industries. Alternative Corporate Governance and Decentralized Media (100:32) Corporate management is experiencing a divergence between Alfred Sloan’s multi-layered bureaucratic hierarchy and Elon Musk’s engineering-centric operational model. Musk’s framework discards the chain of command, relying on five-minute engineering reviews, direct executive intervention at bottlenecks, and intense urgency around "ground truth" to weed out non-believers. Simultaneously, the media landscape has shifted from broadcast gatekeepers to decentralized networks driven by algorithmic feeds, video clips, and direct publishing like Substack. Referencing Martin Gurri’s The Revolt of the Public, Andreessen argues that unconstrained digital communication operates as an institutional x-ray machine, revealing structural incompetence in centralized organizations. While Marshall McLuhan warned that a global village triggers tribal panics, Andreessen views this friction as a modern equivalent to revolutionary-era pamphleteering that permanently disperses institutional authority. While Andreessen remains optimistic regarding technological abundance and open communication, the discussion leaves unresolved whether legacy institutions, political systems, and heavily regulated sectors can adapt to hyper-deflationary pressure, or whether heightened transparency and decentralized media will perpetuate institutional paralysis and social friction.

Marc Andreessen与Charlie Songhurst指出，硅谷的技术创新与风险投资本质上依赖于长周期的幂律回报与高信任的“前沿精神”；在当前人工智能与加密货币的技术浪潮中，其底层效用正在打破传统泡沫模式，并重构去中心化传播与社会经济结构。 创投生态的本质：幂律回报与声誉机制 (11:55) 风险投资在机制上不同于公开市场的定投策略，其核心受幂律法则主导。Marc Andreessen指出，投资的成功与注入的资金量几乎无关，单一优质标的的超额收益能够覆盖全部亏损，例如Andy Bechtolsheim早期向Google开出的10万美元天使投资支票，最终换来了约30000倍的回报。在创投领域中，错失巨大成功的代价（第二类错误）远大于亏损本金（第一类错误）。由于VC回报存在高度的跨期随机波动，顶级机构的单期基金甚至可能在15倍回报与保本之间剧烈震荡，因此耶鲁大学捐赠基金主管David Swensen等资深LP通常倾向于维持20至50年的超长周期配置，以避免在周期低谷时错误退出。 硅谷的核心竞争力根植于一种基于长期声誉的高信任生态。与好莱坞那种“一部电影立项就挤占另一部机会”的零和博弈不同，技术行业历史上具有正向乘数效应，促使参与者维持互助与建设性合作。顶级VC为早期初创公司提供“信誉过桥贷款”，帮助其在尚未自证的情况下吸引顶尖工程师与下游资金，从而启动“马太效应”的正向循环。Andreessen认为，硅谷的崛起并非不可复制的地理巧合，而是绝对契约法、深厚资本市场与自选式“边疆开拓精神”的结合，这也是波士顿等早期科技中心在20世纪90年代逐渐落后于西海岸的重要原因。 技术泡沫的辨析：网络泡沫滞后与AI现实效用 (39:00) 辨别泡沫破裂的时机在现实中极难实现。Andreessen回顾2000年互联网泡沫时指出，那场危机并非软件泡沫，而主要是电信与物理基础设施的杠杆与信贷泡沫，并且其破裂是一个在2000年至2005年间经历多次下挫的漫长过程。互联网泡沫发生的核心原因在于基础设施严重超前于终端用户的使用条件：2001年大多数用户仍依赖56K拨号上网，家用宽带直至2005年后才普及，移动宽带更推迟到2012年前后。如果当时宽带普及能提早五年，互联网泡沫甚至不会以崩塌的形式出现。 相比之下，当前的AI热潮在底层逻辑上与dot-com截然不同。Andreessen认为，现代AI并非传统软件的延伸，而是计算机工业80年来从冯·诺依曼架构向神经网络的首次底层重塑，类似于开启了“计算机工业2.0”。今天的AI具备直接且立竿见影的效用，用户对ChatGPT等产品的消费具有实际粘性，并非仅靠资本概念催生。Stripe的数据也佐证了这一势头：其平台上排名前100的AI初创企业达到100万美元营收的速度比当年的SaaS公司快了整整四个月，且《福布斯》AI 50强企业中超过四分之三均采用Stripe进行订阅和计费。 双轨经济的割裂：通缩型效率与通胀型垄断 (64:16) 技术进步与AI的普及正在将现代经济清晰地切割为两个平行的世界。以软件、电子、媒体和轻工业为代表的高生产力部门呈现出持续的超通缩趋势，边际成本迅速下降；然而在住房、教育和医疗等领域，价格却在大幅通胀。Andreessen指出，这种结构性失衡的根源在于政府对后者的供给端施加了严格限制（如住房开工审批、医生执照名额限制和大学认证壁垒），同时又在需求端持续注入学生贷款和购房补贴，导致基本生活成本持续攀升，形成了社会性的焦虑与政治撕裂。 此外，鲍莫尔成本病进一步加剧了这一扭曲。高生产力部门推高了社会整体工资预期，但低生产力部门往往借助工会力量和行政手段人为阻碍自动化以维护职位。例如美国港口工人在长期劳资协议下甚至出现过“在职闲置人员与一线工人一比一”的低效现象，波士顿市议会也曾试图通过抵制Waymo自动驾驶来“保住驾驶岗位”。Andreessen预测，若未来AI全面普及，极度高效的技术或将引发类似第二次工业革命期间的生产性通缩，使得名义商业服务成本大幅归零，但社会实际购买力却会因此跃升。 金融与组织的范式重塑：稳定币与极端求真管理 (78:47) 在金融科技领域，传统VC因意识形态和认知偏见错失了加密货币的早期红利。Charlie Songhurst与John Collison指出，传统FinTech一直受限于各国的本地监管与银行体系，呈现碎片化特征，而稳定币（实物资产上链）正在打破这一地缘屏障，年供应量以40%至50%的速度复合增长，为全球可编程货币提供了底层网络。Shopify在结算通道中全面上线稳定币支付，标志着加密应用正跨过类似早期互联网拨号时代的基建门槛，迈向主流实用化阶段。 在微观组织管理层面，传统自上而下的企业科层制正暴露出严重的滞后性，多数濒死企业因回避直面竞争而选择长达数年的“歌剧式慢性死亡”。与此相对，以Elon Musk为代表的管理范式展现出极高效率：彻底越过管理层直接对话一线工程师、每周直接空降核心技术瓶颈、以及将核心指标极端简化为具体物理量（如SpaceX的“每公斤入轨成本”或Tesla的“周交付量”）。这种“不惜一切代价追求基本事实”的文化虽具争议，却通过缩短反馈周期和剥离繁文缛节，对传统大企业病构成了降维打击。 媒介形态与话语权力：去中心化曝光与全球信息孤岛 互联网内容分发正在瓦解中心化机构的权威。传统电视与报刊的话语体系被社交媒体、Substack以及短视频剪辑彻底解构，短视频片段在X或TikTok等平台的传播量通常是完整节目的上千倍。Andreessen援引Martin Gurri在《公众的反叛》中的观点指出，新媒体环境不仅是信息的放大器，更是制度的“X光机”，它彻底暴露了传统巨头和官僚体系原先依靠公关包装隐藏的真实缺陷，使得中心化权威的信誉面临系统性瓦解。 这种演变复刻了18世纪殖民地时期北美匿名小册子作者激烈论战的革命性语境，但在全球算法的推动下演变为马歇尔·麦克卢汉所预言的“全球村”。当前单一条短视频可在数小时内跨越语言阻隔引爆全球舆论，使得整个人类社会仿佛置身于全景敞视的传统村落之中，容易滋生普遍的情绪化、集体恐慌乃至社会功能失调。尽管如此，Andreessen依然持乐观态度，坚信不受滤镜遮蔽的自由表达与透明化竞争最终将推动社会走向自我纠偏。 对话表明，硅谷的历史动力在于承担技术不确定性与抵御短期市场情绪。然而，随着AI加速普及与新型媒介进一步瓦解传统权威，前沿创新在面临生产力爆发的同时，也将不可避免地遭遇受监管部门的阻力与全球舆论场泛情绪化的严峻考验。

[8:37] Marc Andreessen: B2B now stood for 'Back to Banking,' and B2C stood for 'Back to Consulting.' [19:02] Marc Andreessen: When the company goes bankrupt, at least it ends... When you pass on the company that succeeds, the pain is forever. [22:33] Marc Andreessen: You're lucky if your friends knife you in the chest [27:07] Marc Andreessen: You flush all the status seekers, you flush all the tourists [45:58] Marc Andreessen: I'm not talking about building a genius computer brain. I'm talking about building a mediocre computer brain, like the president of AT&T. [66:41] Marc Andreessen: Everything that's optional and fun is getting super cheap, and everything that's actually necessary to raise a family is getting like hyper-expensive. [68:25] Marc Andreessen: If there's a hole in your drywall, it's cheaper to put a flat-screen TV over it than it is to repair the drywall. [79:32] Chris Dixon (quoted by Marc Andreessen): What nerds do on nights and weekends. [89:57] John Collison: People would much rather lose slowly over five years, than have the conversation that involves a dramatic change to stop losing. [95:40] Marc Andreessen: Don't. If you need to do that, sell your company.

[0:00] Do you mind if I start with a couple of questions? [0:02] I mean, sure. [0:05] Cheeky means what exactly? [0:06] Okay, in the context of a cheeky pint, [0:09] it is a pint you're not really meant to be having. [0:11] And when it becomes established, [0:12] it starts to attract establishment people. [0:14] The Social Network problem. Right. Exactly. [0:16] And in that sense, the downturns [0:17] as much of a pain in the butt as they are, [0:19] are probably helpful. Are good, because [0:20] you go back to banking, you go back to consulting. [0:22] Yes. I'm sorry, why is there [0:23] more of a risk-taking on the West Coast [0:24] versus the East Coast, because like...? [0:25] Ah, the fontier. [0:27] because FOMO leads to high trust. [0:30] That has a cynical truth to it. [0:32] Category-two errors are much, much worse. [0:34] By the way, they torture you for fucking decades, [0:36] because you read about [0:37] the success cases that you've screwed up all the way up, [0:40] and so, you just learn the hard way, [0:42] you have to be extremely open-minded. [0:44] I have found people willing to tolerate [0:45] any level of chronic pain in order [0:47] to avoid acute pain. Avoid acute pain. [0:49] People would much rather lose slowly over five years [0:52] than have the conversation [0:53] that involves a dramatic change to stop losing. [0:55] All right, there we go. Ha. [0:56] All right, very good. [0:58] Anyone need anything else? [0:59] Finally, a legitimately Irish bartender. [1:02] I have a scheduling issue with these, [1:03] because 5:00 pm clearly after-work pints, acceptable. [1:07] 4:00 pm, I don't know, after work, [1:09] if you're a banker or whatever. [1:11] 3:30 pm, like now you're just drinking [1:12] at the office, you know? [1:13] Marc Andreessen has been around the internet [1:15] since the very beginning, really. [1:16] He co-founded Netscape. He invented the image tag. [1:18] He was there at the beginning, [1:20] and later, he co-founded the venture capital giant [1:22] Andreessen Horowitz, so I'll be speaking to him, [1:24] along with our mutual friend Charlie Songhurst. [1:26] Cheers. Cheers. [1:27] Good to see you guys. [1:30] Do you mind if I start with a couple of questions? [1:33] I mean, sure. [1:36] Well, there's just a couple of things. [1:37] As a Midwestern-American boy, [1:39] there's just a couple of things that, [1:40] this is not my natural habitat. [1:42] Okay. Cheeky means what exactly? [1:43] Okay, in the [1:45] context of a cheeky pint, [1:46] it is a pint you're not really meant to be having, [1:50] and so, if you were meant [1:52] to be going home right after work, [1:54] and instead you stole away with a few coworkers, [1:58] just off the books, aren't meant to be at the pub right now, [2:02] that would be a cheeky pint. [2:03] And then the thing about pint that's just really puzzling [2:05] is that everything else in Europe is like, [2:09] it should be the cheeky deciliter. [2:10] I see. Right? [2:11] And so, why is pint used with reference to alcohol, [2:14] but not with actual measurement? [2:16] Because, Guinness and alcohol generally, [2:21] is part of a rich tradition. [2:22] Guinness dates from the 1700s. [2:24] It's the reason we have the canal system in Ireland. [2:27] It was the largest company in Ireland at one point. [2:30] Often, the longest-tenured institutions [2:34] are universities and breweries, [2:37] and you look at the Belgians and things like that, [2:39] and so, I think tradition survives better in alcohol [2:42] than it does in real sense. [2:43] Okay, I have several more questions, [2:44] but I will suspend them for the [2:46] purpose of this conversation. [2:48] Okay, I like this new format [2:49] that we're inventing. [2:49] So, where I want to start is we have here [2:52] various bits of Marc Andreessen memorabilia [2:55] and a still from one of my favorite pieces [2:57] of Marc Andreessen content, [2:59] your Miller Lite commercial. [3:00] Oh, that was so much fun. And it was only in a rewatch [3:02] that I realized it was with Norm McDonald. [3:04] Norm McDonald. [3:05] What was it like meeting him? [3:05] The one and only. [3:07] My experience, comedians are always [3:08] a little bit interesting to meet, [3:09] because they're professionally funny, [3:11] and so, they're interested in being [3:12] interpersonally funny is not that high. [3:15] Oh. Ah. [3:16] Yeah, yeah. because it's a lot of stress [3:18] and pressure, I think. [3:19] I see. I mean, [3:20] he was very naturally funny. [3:21] Yeah, yeah, yeah. [3:22] But— But he wasn't always on. [3:23] He was not always on, [3:24] and I will tell you in context, [3:25] you will see in the commercial, I just say in context [3:27] it looks like we were in the coolest nightclub in the world. [3:29] I will tell you it was in the middle of the day [3:32] in what they call the Inland Empire in LA in some warehouse, [3:38] and was— It was not as cool [3:39] as it looks. It was like [3:40] 110 degrees outside, it was like 130 degrees inside, [3:44] there was no air conditioning, [3:45] because it would screw up the sound, [3:46] and then to create the smoky-nightclub effect, [3:49] they spray vegetable oil. [3:51] Ooh. [3:52] Not water, vegetable oil. Not water, vegetable oil, [3:54] because it has to— [3:55] Oh, it has to actually create a— [3:57] It has to linger. Yeah. [3:58] It has to linger. [3:59] The director was great, and he was tremendously tolerant [4:00] of me with no actual experience doing anything like that, [4:03] but I think he did think [4:04] he was Stanley Kubrick, [4:05] because we did like 150 takes. [4:08] He was really into his Miller Lite ad. [4:09] Yeah, and so, like hour six [4:11] of nearly passing out from the heat [4:13] and choking on vegetable oil was not— [4:15] I see. The most... [4:17] Oh, and then the other kind great claim to fame is, [4:20] it was a week later, Miller fired their ad agency, [4:23] which I would like [4:25] to think that I— [4:26] Was causal. [4:27] I bear some responsibility for [4:29] having no acting ability, yes. [4:30] Okay, so, the thing I want to get into you guys about, [4:35] or spend a lot of time on, is the history of the valley. [4:38] One interesting place to start might be [4:40] can you tell when you're in a bubble? [4:42] So, my experience is no, [4:44] and the nuance that I would put on that, [4:48] I'll describe two reasons. [4:49] Number one is there's an old line [4:50] with respect to economists, [4:51] that also applies, I think, to investors and entrepreneurs, [4:54] which is economists have predicted [4:55] nine of the last two bubbles, [4:57] or nine of the last two crashes, [4:58] and so, it is extremely common. [5:01] It's a difficult question, [5:01] because it's extremely common for people [5:02] to call it a bubble. [5:04] When they're correct, they will then go around for years [5:07] claiming that they're the one who called it. [5:08] What you find with those people generally, [5:10] is they were calling it continuously [5:11] for the 20 years earlier. [5:13] Peter Singer. Peter, for example, [5:15] or earlier, there's a famous Barron's. [5:19] Barron's it's still around, [5:20] but it used to be extremely important investing publication. [5:23] There was a columnist for Barron's, [5:25] something Abelson, Alan Abelson, [5:26] and literally, he wrote the same column for 40 years. [5:30] "The end is here." "It's all going to crash." [5:31] "It's all a giant bubble," [5:32] and he wrote that, I think continuously, [5:34] from, I forget the exact years, [5:36] but from like 1975 to like 2015, [5:39] and so, you have this kind of Cassandra thing, [5:43] where they kind of dine out on it, [5:44] and so, I find generally, that those kinds of people [5:47] don't have predictive ability. [5:48] And then I will tell you, look, the most sophisticated [5:50] hedge fund managers in the world, generally, [5:52] if you look in their backgrounds, at some point, [5:54] if they thought they were in the macro business, [5:56] they will have tried to make the trade [5:58] based on what they view as obviously, a bubble, [6:00] and there were extremely sophisticated [6:02] hedge fund investors. [6:02] They went short tech stocks in the fall of '99 [6:04] and then realized they were wrong [6:06] and then went long tech stocks in Q1 of 2000. [6:08] Druck has talked about it publicly. [6:10] He's talked about it, but there are many others. [6:11] Well, there's another guy who I won't name, [6:14] who's very active today, who's very smart, [6:15] and I was talking to him on the phone about stuff, [6:17] and he just started laughing [6:18] and he said, he's like, "All I know is whenever I think [6:20] the stock market's going to go up, it goes down [6:21] and vice versa." [6:22] And this is a guy who's an investing legend. [6:25] It's something— Was it obvious [6:26] when the bubbles started to burst? [6:28] When was it obvious, in retrospect, [6:29] that that was a bubble? No. No, no, no, no. [6:31] So the— So, is it 2000, 2001? [6:33] Is it only 2004 when you look back? When does...? [6:36] No, so the sort of cliche, which is correct, [6:39] is the market climbs a wall of worry, right? [6:41] So, what happens is, when the market is rising, [6:44] every step of the way, there's some panic attack [6:46] going on about it's immediately going to collapse, [6:49] and then what happens is there are drawdowns- [6:51] Mm-hmm. Yeah. I'm sure you guys have seen. [6:52] The drawdown charts are really [6:54] fascinating to see. Yeah. [6:55] because you— There's a big one in 1998 [6:56] with the Asian crisis. [6:57] So, we all thought that was it. [6:59] This is exactly where I was headed, [7:00] so yes, there was a blow up in '98, [7:02] there was a international crisis, [7:04] and then there was a collapse [7:05] of a big hedge fund at the time called— [7:07] LTCM. Yeah Long-Term [7:08] Capital Management. I read that book recently. [7:09] It was really good. It's a fantastic book. [7:11] Yeah. [7:14] It's a great lesson in [7:15] do not name your hedge fund long term. [7:17] I thought the lesson was don't run 30-times leverage [7:19] on the one trade. [7:20] Oh, there is that, [7:21] and also assume that academic superstars [7:24] necessarily have a feel. [7:25] Mm-hmm. Yeah. [7:26] But yeah, that was it, that's it for IPOs, [7:30] it's over, that's it, the whole thing is going to cave in. [7:32] So, every step of the way, [7:33] and then conversely, we all got so used to it rising [7:35] that there was a lot of speculation. [7:38] I would say the median view among smart people, [7:39] and when the Nasdaq first cracked [7:41] in around March of 2000, [7:43] was, oh, it's just another one of these momentary blips, [7:45] and then the way I remember it, [7:46] we'd have to look at the chart, [7:47] but the way I remember it is, fundamentally, [7:48] that from 2000 to 2005, there were five discrete moments [7:51] where it fell apart. [7:52] It kept cascading down, [7:54] and my favorite version of the story is [7:55] we took our company LoudCloud public in September of 2000, [7:58] and while we were on the road, [8:00] we were on the road for three weeks, [8:01] and while we were on the road, [8:02] the Nasdaq fell in half, [8:03] but that was just one of those things, [8:05] and so, the answer to your question is, [8:08] I'll put it this way, by 2003, 2004, [8:11] you knew that it was really bad, [8:13] and then what are the indicators? [8:15] The indicator that everybody really knows it, [8:17] is the longs all get fired. [8:18] Yep. They lose their money [8:21] and then the PMs actually get terminated, [8:23] and until that happens, there's still, I would say, [8:25] tremendous amounts of either uncertainty [8:28] or you could say, denial. [8:29] One of the great years for owning internet stocks [8:31] was 2003, because you get the bottom [8:33] and then you get this huge uplift, I think in eBay, [8:35] Yahoo, maybe it's 2004. Sure. [8:37] But VC wasn't good through that entire period, up until [8:38] like '07. [8:42] Why is it that public markets is good in '03 and '04, [8:45] but VC just has almost like a lost seven years [8:48] ex Google between 2000 and 2007? [8:50] You could maybe say this, [8:52] you could say the entrepreneurial ecosystem [8:54] got completely flattened. [8:55] Got it. [8:57] By '03, '04, the idea of starting a company was ludicrous. [9:00] Got it. [9:00] And so— So, it maybe created [9:02] too much fear in potential entrepreneurs. [9:03] Yeah, that's right. [9:04] And then, look, the VCs panic. [9:08] They say that's one of the cardinal sins [9:10] you can get into in venture is [9:11] you're actually paying attention [9:12] to what they're saying on TV. [9:13] Yeah. And particularly, [9:14] on the financial news, [9:15] and so, if the Nasdaq cracks, [9:16] it's very hard to keep yourself out of that psychology [9:20] and to be enthusiastic about make an investment, [9:21] but of course, the rational thing to do if you're a VC [9:24] is to keep... [9:25] So, Fred Wilson's the guy who kind of really [9:27] walked me through this originally, [9:27] and his version of this would be, [9:29] yeah, bubbles bust, it's all random and crazy [9:31] and we never know what's going on in the whole thing, [9:32] and you get wrapped up in the psychology, [9:33] and so, his rule of thumb always was [9:36] you have a disciplined mechanical process [9:38] for the pace of investment [9:39] and then also for the pace of exits, [9:42] and you don't deviate from it, [9:43] and a lot of that justification would be precisely so [9:45] that you keep investing at the bottom. [9:48] It's so funny, you see this in the stock market, [9:50] everybody says, "Oh, buy low, sell high." [9:55] Everybody's an expert in bubbles. [9:57] Everybody's read the books, the whole thing, [9:58] but when the market has caved in, [10:02] it is just, well, it's actually really funny, [10:05] because it's negativity. [10:07] It's just overwhelmingly, "You people are idiots," [10:10] like, "This whole thing is stupid," [10:11] "It's never going to recover." [10:12] There's 18 macro explanations for it not to recover, [10:14] and then actually, at the real bottom, [10:16] the other thing I found [10:17] is people just completely stop talking about it. [10:18] Yes. [10:20] The idea of startups— [10:21] Crypto markets are a case study in [10:23] over the three cycles. It's just like [10:24] it never even existed. [10:25] It's the thing you would never bring up at a dinner party, [10:27] and maybe to your point, that's what happened [10:28] with the internet startups in 2003, 2004, [10:31] which is, you would not talk about it, [10:33] if you could possibly avoid it. [10:34] So, in some ways, the social status of internet startups [10:36] in '03 is similar to crypto in 2020. [10:39] Yeah, so, the great joke of that time [10:42] was, the two great VC trends, startup trends, [10:47] of the late 90s were so-called internet companies, [10:49] but B2C— Yeah. [10:50] Business to consumer, and then B2B, business to business, [10:52] by 2003, the line was B2B meant back to banking, [10:55] and B2C meant [10:56] back to consulting, and so, oh, and then, [10:58] this in turn is why, you'll enjoy this a great deal, [11:01] this in turn is why the employment decisions [11:05] of graduating Harvard and Stanford Business School students [11:08] are such a great indicator. [11:09] Yes. They're possibly the best [11:10] indicator of all, of what's happening in the market, [11:12] because if they go into tech, the market's overblown [11:17] and if they go into banking consulting, [11:18] it's a great time to make VC investments, [11:20] and that maybe been the best indicator [11:22] I've seen the whole time, [11:23] because of the social status aspect. [11:24] Yes. Right. [11:25] I think what you're describing is [11:26] you don't think you're capable of making macro calls, [11:28] so you just have to decide what are sensible areas [11:31] to be investing in over multi-decade time horizons, [11:34] tech startups generally, crypto, American dynamism [11:37] and pick your lane, and then you dollar-cost average [11:39] it to them, and then sometimes there'll be bubbles, [11:42] there'll be crypto 2021 moments, [11:44] but that's fine, because if you put the same dollars [11:47] into these areas, I mean rough numbers, [11:49] but consistently put dollars into these areas each year, [11:52] the winners will more than make up for the years [11:53] where everything was hopelessly undervalued. [11:55] Is that basically your framework? [11:56] I would say that's mostly true. [11:58] What I would modify that is [11:59] it's actually not dollar-cost averaging. [12:01] If you're doing it in the stock market, [12:02] it's dollar-cost averaging. [12:03] If you're doing it in ventures, [12:04] it's not dollar-cost averaging. [12:05] The reason is because if you make [12:06] the right venture investment, [12:07] it doesn't matter how much money you put in. [12:08] Yep. The upside is so great. [12:10] Yep. And if you make the wrong [12:11] venture investment, you lose all the money. [12:12] Well, I think- [12:13] But is that actually true? I think Andy Bechtolsheim's- [12:14] Depends how much money [12:15] you put in, so... Andy Bechtolsheim's 100K, [12:16] I think would be 30,000x. [12:19] What's that, sorry? [12:19] Andy Bechtolsheim's 100K. Bechtolsheim. [12:22] Bechtolsheim's 100k into Google would've been 30,000x— [12:24] Yeah. That pays for a lot [12:26] of other investors. [12:27] In venture capital, it just turns out [12:31] that the amount of money invested [12:32] has almost nothing to do with anything, [12:33] and you're not trying... [12:34] Well, here's another thing, you never in venture [12:36] run a bargain shop. [12:37] Ever, ever, ever. No, I agree with that, [12:39] and you— What you need to do, [12:40] so, I guess the way I would just modify what you said [12:42] is it's just you need to keep investing. [12:44] Yes. Yeah. [12:44] The danger is not investing too cheap or too dear. [12:47] The danger is literally stopping. [12:49] Sure, but sorry, when I was saying dollar-cost averaging, [12:50] it was the fixed amount of money— [12:52] Per year. That could deploy, [12:53] because I think the way people get into trouble [12:54] is 2021 comes along, and they raise some giant fund, [12:57] and that one has very poor returns, [12:59] but if you invest 100 million each year, [13:03] then you'll do pretty well. [13:04] And you could also say this, the smartest LPs, [13:06] so David Swensen who was considered to be the smartest [13:08] portfolio manager for liquid portfolios, [13:11] wrote a book where he goes through the following, [13:13] and he talked about this a lot, [13:14] which basically, is for something like venture, [13:16] you really got to look on it. [13:17] You cannot rationally evaluate venture [13:20] based on a single moment in time, a single fund, [13:22] a single sector, any of that stuff. [13:23] You have to basically, look at it [13:24] over a long enough period of time [13:25] where you wash out the specific effects of what— [13:28] Well to prove that is the inter-vintage volatility [13:33] in any given VC is incredible. [13:35] Right. That's right. [13:36] Which shows so much of it is just— [13:38] Yeah, a top VC firm— Structural time. [13:39] Will have some 15x funds— [13:42] Yeah. And some— [13:43] But it's incredibly stochastic [13:44] because Google's founded in '99, [13:46] so, at the height of a bubble. [13:47] Meta's founded 2004, at the bottom. [13:50] Right. That's right. [13:51] There's no pattern that ties to macro. [13:53] It appears to be almost stochastic. [13:55] You just can't predict. You've just got to keep doing it. [13:57] Yeah, that's exactly right, [13:58] and that's the, I would say, [13:59] that's the core fundamental truth of venture, [14:01] which is really, it's something for people [14:03] with a 20, 30, 40, 50-year time horizon. [14:05] You have to get all the way across the cycles. [14:08] Because what happens otherwise, if you're in LP, [14:09] what happens otherwise is, [14:11] the minute you have a fund that's terrible, you pull out, [14:13] and that's precisely when you should have been going in. [14:16] It's the same behavior on the LP side [14:17] that you see on the VC side, [14:19] and so the smart LPs, what they all have in common is [14:21] when they're making a decision to invest in a venture fund, [14:22] they're making a decision to invest in that fund for the [14:24] next five or six months. So how much of an advantage [14:25] for VC is having good LPs? [14:27] Extremely. Extremely, extremely, extremely. [14:29] And again, this is very predictable. [14:32] What happens is, every time the market is hot, [14:34] new LPs show up and pile in, [14:36] and then when the market declines, they back out, [14:38] and so the firms that have the VCs [14:41] who understand the Switzer model [14:42] are able to sustain over time, [14:44] and able to continue to invest in the downturn. [14:47] Many new VC funds are raised [14:49] in every bull market, from basically, tourist LPs. [14:52] Those tourist LPs are extremely reliably prone to pull out. [14:55] So obviously, that leads to the big question, [14:57] which is how causal are the VCs themselves [15:00] to the outcomes of the companies? [15:01] Like it's the big, big question. [15:05] I have an indirect theory on it, [15:06] but I'll... I definitely should not let the entrepreneur [15:10] answer this question, but... [15:11] I just made an incredible [15:13] strategic mistake. [15:15] This is where it all went south for Andreessen. [15:18] I mean— You can see the look [15:19] on his face already. [15:20] One, presumably VC itself, is very impactful [15:26] because Stripe was just, as a practical matter, [15:29] not profitable for quite a few years, [15:31] and I think that was the correct way to build Stripe, [15:33] and so, like so many companies, you build a bunch of tech [15:36] and Stripe in particular, you build a bunch of tech [15:38] and businesses start adopting it and they start growing, [15:40] so you've two lagged curves. [15:42] One is you have to build all this stuff [15:43] and then businesses start using it, [15:44] and then those businesses grow themselves, [15:47] and we just had Tobi from Shopify here, [15:51] Shopify is now a massive business on Stripe, [15:53] but they weren't when they started [15:54] working with us in 2012, [15:56] and so, it's just the classic R&D thing of you [15:59] do work now for economic payoff later, [16:01] and I think that tends to work well in tech. [16:03] And then with specific VCs, [16:07] I want to talk about the Silicon Valley high trust thing. [16:09] VCs act as a very efficient matching algorithm [16:13] between neophyte founders, such as myself, [16:17] and experienced executives, [16:19] and so you have this incredible talent engine, [16:21] and I think in a weird way, [16:25] it's not about the money, at some level. [16:26] People miss that it's about putting together a team [16:30] in a very short order, to go do this hard thing, [16:33] and I think VCs are actually pretty instrumental in that. [16:36] I'll back in from the angel perspective. [16:38] The single strongest correlation of [16:41] how a company will perform is [16:44] how high status the VC who does the Series A, [16:47] is, within the stack ranking of VCs. [16:49] It is far more predictive, sadly, than [16:51] my own selection or any other variable I can find. [16:54] It's almost deterministic. Yeah. [16:56] And look, some of that is because the top-tier VCs [16:58] can get the best deals. Exactly. [16:59] And some of that- Yes. [17:00] Is self-fulfilling prophecy, [17:01] so here's my analysis, having been on both sides [17:03] of the table, John, mapping what you said. [17:05] My analysis basically, is that if you think about, [17:07] mechanically, what's happening with a startup, [17:09] a startup needs to basically get into a loop [17:11] in which is it's accruing more and more resources [17:13] as it goes, and those resources are qualified executives, [17:17] technical employees, future-downstream financing, [17:20] positive brand momentum, public perception, [17:24] customers, revenue, throw weight in the government. [17:27] All of these resources that you need [17:28] to be able to succeed as a business, [17:30] and so, there's a snowball-rolling-down-the-hill phenomenon, [17:34] which is you're either a snowball rolling down the hill, [17:36] picking up resources as you go, gaining size, and scale, [17:38] and scope, and power as you go, or you're not, [17:41] and you're stuck at the top of the hill as a snowflake [17:43] and you're just not going anywhere, [17:44] and so the question is how do you get into [17:47] this kind of aggregation of resources thing? [17:49] Economists call this, what's the term for the things [17:51] that are at the height of the power... [17:52] Preferential attachment. [17:53] Yeah. But it's— [17:53] A sort of beating of companies. [17:54] It's the Matthew principle [17:57] from the Bible, which is, he who has a lot will get more [17:59] and he who doesn't... [18:00] And so, when a company gets momentum, [18:03] you hear about momentum, [18:04] when a company gets momentum, what it means is [18:05] the next resource that you need [18:06] is preferentially willing to attach [18:08] to your thing as opposed to somebody else. [18:10] That's the mechanical process that drives [18:12] the power-law curve. So that creates [18:13] a chicken and egg question— Correct. [18:14] Which is, does the product create the company [18:17] or does the company gather enough resources [18:19] to create a product? [18:20] Yeah, so that's part of it, but again, [18:21] to create the product, [18:23] it's often now, just a process of, [18:25] it's also, okay, you got to create the engineers, [18:27] and then you got to actually feel the product [18:29] and... [18:30] Let me give you an example. [18:31] You got to have you got to have top-end security engineers. [18:33] There are only so many top-end security engineers. [18:36] Where do they want to work? [18:37] They want to work at the top companies. [18:38] If you're a brand-new startup, [18:39] how do you convince somebody [18:40] that you're going to be a top company? [18:40] You raise money from a top-tier VC. [18:42] So that happens over and over again. [18:43] The prosaic way that I put it is, [18:45] as my experience as a founder is, [18:46] a top-tier VC is a bridge-loan of credibility [18:50] at a point in time when the startup maybe deserves it, [18:52] but just doesn't have it yet, [18:54] and that credibility is harvested in the form [18:56] of primarily personnel, money, and brand, [18:59] and those three things turn out to be [19:01] really important in the beginning. [19:02] We're talking about the Silicon Valley ecosystem here [19:07] and you referenced Andy Bechtolsheim [19:09] and his investment in Google. [19:11] One thing that I find funny about that story is, [19:13] that's the case where he just wrote a 100K check to them. [19:15] He actually wrote a 100K check to Google Inc., [19:17] even though they didn't have a company. [19:18] I think he'd gotten in his Porsche and drove off, [19:20] and is like, "Here you go," [19:21] but there was no terms, there was no nothing, [19:23] and that obviously worked out really well for him. [19:25] But that's not unusual, I've heard other story... [19:27] I think we even got some check like that [19:29] where again, it was just like, "Tell me the terms later," [19:31] and Silicon Valley is very-high trust. [19:34] How did that come about? [19:36] Let me tell you, that story is a great story. That is true. [19:38] And I will tell you there is another part of that story, [19:40] which is the venture firms that turned down Google [19:41] on the Series A, which is the whole other side of things [19:45] that maybe we should talk about, [19:46] because, in retrospect, it all looks obvious. [19:48] At the time, it's not so— [19:49] Sure, but it wasn't obvious. [19:50] Yeah. Not obvious, so maybe that [19:51] reinforces what you're saying, [19:52] which is it's definitely not obvious. [19:53] Look, quite frankly, you could have all kinds of theories [19:57] about this, do all kinds of things. [19:58] Talk about how wonderful everybody is. [19:59] I think the prodigal reality is anybody who's been [20:00] in the Valley for a while has had the experience, [20:03] typically in the form scar tissue [20:05] where there was some kid in a T-shirt [20:06] with some crazy idea, and you were like, [20:07] "Okay, that's great." [20:08] Oh, the opposite? [20:10] Yeah, yeah, yeah. Yeah, you pat 'em [20:10] on the head and they go off on their way, [20:11] and then they turn around and five years later, [20:13] it turns out, "Oops, that was Mark Zuckerberg. [20:15] Shit. I had my moment. I had my chance." [20:17] The problem with missing, [20:19] remember it's category one— So it's basically, [20:20] okay, that's how ventures is, it's FOMO leads to high trust. [20:23] That has a cynical truth to it. [20:26] It feels weird. Like, yeah. [20:26] If you sit around, yeah. [20:28] It goes to category-one versus category-two error. [20:30] Again, it goes back to the economics, [20:31] which is Andy's $100,000, [20:33] if it got stolen, it's, he only loses $100,000. [20:36] If he gets it right, he makes the 30,000x return, [20:38] and so, there's this thing, [20:40] which you learn over time, is the category-two errors [20:42] are much, much worse, [20:44] and, by the way, they torture you for fucking decades, [20:47] because you read about [20:49] the success cases that you've screwed up, all the way up, [20:52] and so you just learn the hard way. [20:54] You have to be extremely open-minded for people— [20:56] I have a confession here, [20:57] which is what I tell entrepreneurs off to see VC, [20:59] I'd say, "Look, don't try and convince me [21:01] you're going to be successful, just try and create a fear [21:04] that there's this possibility for the next 20 years, [21:07] they might regret this— [21:08] That's so painful. As their past [21:10] personal billion that they missed. [21:11] When the company goes bankrupt, at least it ends. [21:14] Yeah. It's over. [21:15] The pain is over . [21:17] When you pass on the company that succeeds, [21:18] the pain is forever. [21:19] It's like the asymmetry of shorting. [21:20] You're shorting [21:21] the entrepreneur. Oh yes, absolutely. 100%. [21:23] It's a horrible mistake, and so, as a consequence, [21:24] there's just this thing of, [21:26] what it leads to is this incredible sense of possibility [21:28] and incredible sense of optimism, [21:30] in a very positive way, [21:31] which is you just need to be extremely open to the idea [21:33] that you're going to run into [21:34] the next big thing at any moment. [21:35] And they say, karmically, you want to really [21:38] put yourself out there to be part of that. [21:39] I think that's true, [21:40] but I think that's maybe a different thing. [21:41] You're describing that success can come from anywhere. [21:44] There's a big asymmetry in success where companies [21:46] can, 10,000x, where they can't go down [21:48] by more than 1x from their present position, [21:52] but it seems like particularly the business culture, [21:55] and even moving outside the fact [21:57] that startups get really big, is particularly high trust. [22:00] So you have all of investing happens based on handshakes, [22:02] and people can just shake hands and it's going to happen, [22:05] and trust that everything happens there. [22:07] Even when we buy companies, [22:09] we generally agree with the founders at a high level [22:12] of the terms and there might be a single-page [22:13] or a two-page term sheet, [22:15] and obviously lots of due diligence will happen after that, [22:18] but it won't be the kind of East Coast [22:21] process, private equity process after that, [22:23] where everyone's trying to pull a fast one, [22:24] and you can't trust the lawyers [22:26] as fast as you can throw them, [22:27] so there seems to me, [22:27] there's a particular kind of high-trust relationship [22:30] in how all the actors work [22:33] with each other on that. I was going to ask, Marc, [22:35] why the East Coast and why Europe [22:37] hasn't to Silicon Valley, [22:38] whereas you have Detroit, but then Japan copies it. [22:42] And I think maybe he's actually already [22:44] answered the question, which is maybe [22:46] because I haven't had those 10,000x returns, [22:48] they haven't instilled the fear of FOMO, [22:51] and it's the fear of FOMO that means you've got to [22:54] take a trusting bet on a new person, [22:58] and maybe that's the kernel [23:00] that creates a high-trust ecosystem. [23:02] Yeah, and then maybe just add, I think maybe you're right [23:04] that I'm being a little bit too cynical in my answer. [23:06] It's also that you want your reputation [23:08] to project into the future, [23:10] and so if you have a reputation, [23:12] it's a fairly close-knit community, [23:14] if you have a reputation for being helpful [23:16] and being positive, and constructive, [23:17] and value-add, then that plays well, [23:19] because then the person you've done something nice for [23:22] is going to introduce you to other people in the future, [23:23] and even you're— It's a very repeat game. [23:24] Right, right, it's the [23:26] ultimate repeating game. [23:27] Yeah. And so there's that, [23:28] and then, look, I think the other side of it, [23:29] that you guys alluded to, but I think is very important, [23:32] which is it's not zero sum. [23:33] When I talk to my friends in Hollywood, [23:34] which is not that far away, [23:37] and is it's own entrepreneurial ecosystem, [23:40] and if you talk to anybody in Hollywood, they're like, [23:41] "Oh, my God, this is a shark tank." [23:43] You're lucky if your friends knife you in the chest. [23:46] Generally, it's in the back. [23:47] It's this constant thing, and the reason is, [23:49] because, at least my analysis, there's a fixed amount [23:51] of money to be spent and made in movies, for example, [23:54] and if my movie gets green lit, it means yours doesn't. [23:57] And so even if we're close friends, [23:58] we're going to undermine each other as much as possible. [24:00] Whereas in tech, at least historically, [24:01] you have this multiplicative generative thing [24:04] where it keeps expanding. [24:05] So why did nowhere else manage [24:07] to get that ecosystem going? [24:09] If you look at the history of this last 50 years, [24:11] one of the stories that will come out is a utter uniqueness [24:14] that tech almost became a Silicon Valley, [24:16] or at least a West Coast monopoly. [24:18] Yep. There's no [24:20] precedent for that in any other industry. [24:21] Well, I think we're back to that. [24:22] Exactly. I think, [24:24] you see this in data actually, already, [24:26] AI was re-consolidating— Yeah, exactly. Exactly. [24:27] into basically two places on Earth, [24:29] and only one in the west. [24:31] No part of the industrial economy had that dynamic. [24:33] Right. What is it? [24:33] So, there have been a long parade of officials [24:37] from other cities in the US [24:38] and from other countries who have come to the Valley [24:39] in the last 30 years. [24:40] I've met with many of them. [24:41] They all ask that question. [24:42] I answered as follows, which is, [24:43] there are a set of things that you need all in combination, [24:46] and then they usually, at that point, [24:48] they get a stricken look on their face [24:49] and they say, "Well, what if we can't do [24:50] any of those things?" [24:50] Yeah. And so— [24:52] What if we build a really linear city? [24:54] Exactly. [24:55] Well, actually, it's surprising the number of people, [24:57] and I don't want to badmouth people, [24:59] because people should try to make these things work [25:01] and I'm proud of them for trying, [25:02] but literally the number where it's like, [25:03] "Wow, if we just built the right buildings... [25:06] Yeah. "This would happen." [25:08] That's actually fairly common, [25:09] and anybody who's been to Silicon Valley knows, [25:10] the key factor— Exactly, yeah. [25:11] The key to it is not- Go on El Camino Real, [25:12] it's not the buildings. [25:13] It is definitely not the buildings. [25:14] Yeah. And so I think [25:15] it's a formula, and I think it's a list of things, [25:17] and it's like making a cake. [25:18] They all have to be in the cake, [25:19] and the best way, I think I can describe it is, [25:22] it's a set of things that have to do with stability, [25:24] and maturity, and rule of law. [25:26] So you need absolute contract law. [25:28] You need liquid-deep capital markets. [25:30] You need expert specialists in all these different areas [25:33] that have real experience, accounting and everything else, [25:36] and so there's a maturity and a depth, [25:39] and it's that stuff that [25:40] developing-market countries struggle with, [25:42] but at the same time, you need the Wild West [25:45] and you need the spirit of adventure, [25:47] and the craziness, and the willingness to take risks, [25:50] and if somebody fails— And that's what [25:51] the East Coast missed? [25:52] And that's what the East Coast missed, [25:53] and that's what Europe doesn't get right. [25:55] At least, when I talk to my friends in the East Coast, [25:56] or my friends in Europe, that's what they say, [25:57] like, "Well, I can't take that kind of career risk. [26:00] That's crazy," and look, in a lot of countries [26:04] and in a lot of cultures, if you take a risk like that [26:06] and it doesn't work, it's a real problem. [26:08] And sorry, why is there more risk taking [26:09] on the West Coast versus the East Coast? [26:10] because like— Ah. The frontier. [26:13] There's no established hierarchy. [26:15] The frontier. The frontier. Really? [26:16] The frontier. [26:17] It's the frontier. Yeah, if you were in 1985, [26:17] you could go join— It's all in, [26:19] what's his name? [26:19] The Winslow, "The Frontier" guy from— [26:22] I was going to say it's in "Bonfire of the Vanities" too. [26:23] Okay, true. You would go join [26:24] Goldman Sachs, you would join McKinsey. [26:27] You would join existing institutions [26:29] and go up them on the East Coast. [26:30] Those just didn't exist on the West Coast. [26:32] You effectively had a country of [26:33] 50 to 70 million people— But don't worry. [26:34] With no hierarchy— There was Wells Fargo, [26:35] there was lots of institutions [26:37] that you could join. Yeah, but were they [26:39] prestigious enough that they trapped young talent? [26:44] Another way to say this is, why did Stanford [26:47] do so much better than Harvard and MIT? [26:49] Because obviously, the input quality is the same, [26:52] so there has to be something in the place [26:54] they're sitting that creates a difference. [26:55] I think there's a frontier spirit. I really do. [26:57] So it— [26:59] But you're always skeptical of cultural explanations [27:01] in other places. [27:01] Well, there's clearly a talent-aggregation effect. [27:03] So there's clearly talent-aggregation effect [27:05] that takes place inside the US. [27:06] And look, most of the great people in Silicon Valley [27:07] did not grow up in Silicon Valley. [27:10] My wife grew up here in Palo Alto. [27:11] I tease, I call her a townie. [27:13] Right? [27:14] By the way, she has three more degrees than I do, [27:16] so it's definitely not a status thing, [27:18] but most people are imports. [27:19] They get imported all through the entire rest [27:21] of the country, and around the rest of the world, [27:22] and so it's definitely a selector, [27:24] an attraction point for talent, and that's a big part of it, [27:27] but, look, I think if you just trace the history, [27:31] it's not an accident that both Silicon Valley [27:32] and Hollywood are the places that they are [27:35] because the people involved went west as far as they could [27:37] before they were literally stopped by the Pacific Ocean. [27:39] Right? [27:40] It was the ultimate selector in the build out of the country [27:43] to the people who were the most oriented towards risk, [27:45] and, to your point, independence, [27:47] and doing their own thing, [27:48] and that was true in the gold-rush days in 1850 [27:53] where San Francisco was ground zero for that. [27:54] It's equally true today. [27:55] Hollywood is the exact same thing. [27:57] In Hollywood's case, it's actually funny, [27:58] because one of the reasons they needed [27:59] to get so far away is they were trying [28:00] to evade Thomas Edison's patent enforcers, [28:04] because Thomas Edison owned the patent for the film cameras [28:06] and the original Hollywood entrepreneurs [28:08] had no desire at all to pay for that, [28:09] and then Edison would hire the Pinkertons [28:11] to come bust up the movie sets. [28:14] But you see what I'm saying. [28:16] Rogue, renegade, iconoclastic— [28:18] How about tech? But do you think the [28:19] inverse is true? [28:20] Do you think that certain people [28:21] didn't move because it wasn't a fun city [28:24] that had hit the scale of London [28:26] or Europe? Oh, 100%. Yeah, yeah. [28:27] No, look, we all have lots of friends in New York [28:30] and London, and they're all just like, "Wow." [28:33] My friends in New York, I don't know, [28:34] if you get two pints of this into them, [28:36] they'll be like... [28:36] They literally don't understand [28:37] why anybody doesn't live in New York. [28:38] Yeah. Well, I think they'll [28:40] tell you that at 9:00 AM [28:41] on a Monday morning. [28:42] You don't need to get any drink into them. [28:43] That's a very good point. [28:44] It's a New Yorker cover. [28:46] Yeah, essentially. I trying to— [28:47] So, the frontier and a mining camp, [28:48] you have to be willing to move to the mining camp? [28:50] I think so. [28:50] And then this gets into the danger. [28:53] This is like the back-to-banking [28:53] and back-to-consulting thing. [28:55] The danger, in a lot of ways, [28:57] is it becomes established, [28:58] and when it becomes established, [29:00] it starts to attract establishment people— [29:01] The Social Network problem. [29:02] Right. Exactly. [29:03] And in that sense, the downturns, [29:05] as much of a pain in the butt as they are, [29:07] are probably helpful. Are good because [29:09] you go back to banking, you go back to consulting. [29:10] Yes, and the only people who are left, [29:12] and by the way, this was Silicon Valley [29:13] when I arrived in '93, this had happened, [29:15] and then this was Silicon Valley in 2004 as we discussed, [29:18] which is, you flush all the status seekers, [29:20] you flush all the tourists. [29:21] It's like fuel management for fire. [29:23] Correct. Exactly. [29:24] 100%. You clear out the brush. [29:27] Now look, how long can this last? I don't know. [29:30] We're in a country that has, at least certainly [29:32] over the last 60 years, [29:33] has had a strong tendency towards stagnation. [29:35] The thing that has kept this whole thing going, I think, [29:37] is just that there are these new platforms, [29:40] these new paradigm shifts in technology. [29:41] Everyone loves the defense-company explanation [29:44] for Silicon Valley— [29:45] That's a big part of it. Yeah. [29:46] That's been part of it. Yeah. It's real. [29:47] So, Steve Blank has done the best reconstruction of this. [29:50] The typical Silicon Valley history [29:51] goes back to the 1950s with HP [29:53] and the 1960s with the chip companies. [29:57] I think he makes a very compelling case. [29:58] The real history was actually defense-tech startups [30:00] in the late 1920s, 1930s, [30:02] and you still see remnants of that [30:04] if you drive around Sunnyvale. NASA Ames. [30:06] NASA Ames, but this is the place where early, [30:08] I forget the exact, but early radar [30:10] and early missile-guidance systems and all that stuff, [30:13] avionics, a lot of that was innovated here [30:14] in the exact same way, [30:15] and that was 100 years ago. So, if you could go back, [30:17] could you AB test it? [30:18] Is there any way you could have made Silicon Glen [30:21] or whatever the Boston corridor was called, successful? [30:24] Well, they did. [30:25] And keep it successful versus the Valley. [30:27] That's the problem. [30:28] Was there a point where it could have gone the other way, [30:31] or was it inevitable from the 50s? [30:33] Like you're in 1970. Can it go both ways still? [30:36] So, when I arrived in the Valley in '93, [30:38] I think it's fair to say the Valley and Boston [30:39] were probably considered neck and neck— [30:41] Mm? Yeah, and sort of— [30:42] That late? In Boston, [30:44] these are kind of forgotten now, [30:45] but DEC- Mm-hmm, oh yeah. [30:46] DEC in digital, it was like— [30:47] Ashton-Tate. Extremely important company. [30:49] Ashton-Tate the inventor of the word processor, [30:51] and I think, was there Lotus— [30:52] Lotus was. Yeah Was there, 1-2-3 [30:54] was there, and then you had, later years, [30:57] other great companies, EMC and others, [31:00] and then there's a great book [31:01] called "Soul of a New Machine," [31:02] which is one of the great all-time startup books, [31:04] which is about a supercomputer company in Boston [31:07] in the late 80s. [31:08] It was just extremely excellent literary book, [31:11] and it really tells the story of a startup, [31:13] but it also tells the story of Boston [31:14] in that time and place, [31:15] so a lot of leading-edge supercomputing stuff was there. [31:17] By the way, Thinking Machines was there. [31:19] The original [31:20] supercomputer company. Wow. [31:21] The original computer. [31:22] The original. Yeah, exactly. [31:23] Yeah, so Danny Hillis, the company that's the forerunner [31:27] of what we think of today as large-scale AI [31:29] and cloud stuff, was there, [31:31] and look, MIT was there [31:34] and generated huge numbers of smart people, [31:38] and so it worked really well for a long time, [31:40] and then basically, in the mid 90s, it separated, [31:43] and then people in Boston will say that again, two pints in, [31:46] they'll say that the final blow [31:48] was probably when Mark Zuckerberg could not raise [31:49] venture capital for Facebook, [31:51] and had to leave and come west. [31:53] That was a meaningful signal? [31:54] That was sort of [31:56] the last moments. We can call it [31:57] the chapter marker. [31:58] Yeah, it was just like, [31:59] "Okay, if we couldn't do that one," [32:00] and then, by the way, in the counterfactual, [32:03] had he stayed in Boston, [32:04] maybe there would be an entirely new [32:06] ecosystem there that doesn't exist today. [32:08] Yeah, so I think, basically, it worked for a while, [32:10] and again, this is why I locked in on frontier spirit, [32:12] so what Boston has is all of the stability aspects [32:16] that we were talking about. [32:17] Okay, so if that— They just didn't [32:18] have the same frontier spirit, [32:19] and it just turned out that, [32:20] put it back to preferential attachment, [32:22] it just turned out on the margin, [32:23] the smartest people from MIT wanted to come here, [32:26] and that was basically it. [32:27] So, if that's how to move ecosystems, [32:29] same question, but for companies, [32:30] what's the company that could have been a trillion [32:33] that didn't, that you would have to change the least [32:35] to make it a trillion? [32:36] You know, they get that one exec, [32:37] they get that one lawsuit, that just goes differently. [32:41] There's many, many, many. [32:42] The all-time story of that is a company [32:44] called Digital Research, [32:45] which should have been Microsoft. [32:46] Yeah. And there's a famous, [32:48] I can tell the whole— [32:49] Please. Oh, oh yeah. Okay, okay. [32:51] So, the story roughly goes as follows. [32:53] It's in the books, but it roughly goes as follows, [32:55] which is, so Bill Gates and Paul Allen [32:57] had this little software company, originally Albuquerque, [32:59] down the street from "Better Call Saul," I imagine, [33:02] which they moved [33:03] to Seattle, and they were building, very early, [33:05] they were building programming tools for computers. [33:08] And so, when I first used Microsoft as a kid, [33:10] it was Microsoft Basic. [33:11] They were a compiler company, [33:12] or interpreter company, not an OS company, [33:16] and then there was this PC wave [33:18] with all these, basically, [33:19] these kind of cat and dog early PCs from '76 to '82, [33:23] and they basically sold the basic interpreter [33:24] to all those companies, and that's how they got going, [33:27] but they weren't in the operating system business. [33:28] And then IBM decided, famously, [33:30] to enter the PC business, [33:31] and then there was a network connection [33:34] with Bill Gates's mother and the CEO of IBM, [33:35] and they were on a board together, [33:36] and it resulted in the IBM team coming out [33:39] and going up to Seattle [33:40] and buying a license to Microsoft Basic, [33:42] which is what everybody did in those days, [33:43] and then the IBM team asked Bill Gates, [33:45] like, "What operating system should we use?" [33:47] And he's like, "Oh, well, the standard operating system [33:48] for PCs is called CPM," which at the time was true, [33:51] it was the standard operating system for early business PCs, [33:54] and they said, "Well, who makes that?" [33:55] And he said, "Well, there's a company [33:56] called Digital Research down in Santa Cruz in California. [33:59] There's this guy Gary Kildall, you should go see him." [34:01] And this was the synergistic relationship that he had [34:03] with Digital Research at that time. [34:05] So the story goes, the IBM team, [34:07] which is 20 lawyers in blue suits, [34:09] get on a plane, go to Santa Cruz, [34:10] they show up at the office to meet with Gary Kildall, [34:12] discuss licensing CPM, and Gary Kildall, [34:15] being a frontier-like person, [34:17] decided not to come to the meeting, [34:18] decided he'd rather go flying [34:20] that day, John. [34:22] That's a reasonable thing to want to do. [34:25] And instead, had his wife, [34:26] who was the company's general counsel, negotiate the NDA. [34:29] IBM was famous for its lawyers, and the lawyer was not about [34:33] to sign the NDA and the day ended inconclusively, [34:35] and the IBM team was like, "All right, this is ridiculous," [34:37] and they went back up to Seattle, [34:39] and they told Gates, "If you can't find us [34:40] an operating system, the deal for the interpreter is off," [34:44] and Bill said, "Give me a few days," [34:46] and Bill literally went down the street [34:48] to an independent developer named Tim Patterson, [34:51] licensed what at the time was called QDOS, [34:53] quick and dirty operating system, [34:55] which is the true name of DOS [34:57] and, for a $50,000 flat fee, [35:01] turned around and sold it to IBM, [35:04] that created MS-DOS. [35:05] The kicker to the story is, 30 years later, [35:07] Gary Kildall was knifed to death in a bar fight. [35:10] Oh, my God. Yes. [35:11] Oh, God. Sorry. Sorry. [35:12] Something changed terribly there. [35:14] I didn't want to bring the room down. [35:15] That was... But it should have, [35:17] again, counterfactual and who knows, who knows, [35:19] who knows, but you know. [35:20] But I think, it seems hard to argue [35:22] that Digital Research would've become [35:23] a trillion dollar company because Bill Gates [35:26] had such a killer commercial instinct [35:28] that there were, I mean obviously, the IBM OS moment [35:32] was the biggest moment, [35:33] but there were several other moments in Microsoft's history [35:36] where they steered things, and it doesn't feel like [35:38] if they get the IBM OS pick, that then you magically [35:43] become a giant company. Oh, no, no. [35:44] Definitely, you don't magically become the giant company, [35:46] but again, this goes back to preferential attachment. [35:49] That IBM deal's a classic. [35:51] It's impossible to remember how important IBM was [35:53] at that time. Yes. [35:54] IBM in the mid 80s was 80% [35:55] of the market capitalization of the entire tech industry. [35:57] Mm-hmm. [35:58] They were the absolute gorilla, [36:00] and by the way, the IBM PC, [36:03] and then the clones ultimately, [36:04] that came out of it, completely standardized [36:06] the industry, but all of the PC companies [36:10] from before that went away. [36:12] Yeah. It was an extinction-level [36:13] event for everybody else, [36:14] and so whoever got that deal, [36:16] had he not gotten that deal, [36:17] it's not even clear Microsoft would've stayed in business. [36:19] Now, having said that, he gets obviously credit [36:21] for everything as well. There is a trend where [36:22] if you go to the absolute cutting edge of tech, [36:24] they're so sort of wilderness people, [36:27] that they don't have the conscientiousness. [36:29] Correct. Right, yeah. They go flying instead of [36:31] turning up to the meetings. [36:31] That's right. And it's almost like [36:32] you get a second generation who go to the frontier, [36:34] but a conscientiousness enough to institution build, [36:37] and those become the super-big companies. [36:38] Yeah, by the way, Dell's another classic case dated [36:41] from that same time. [36:42] Dell computer was founded at the same time. [36:44] There's like 400 IBM-clone companies at that time, [36:47] that were actually in the process of going under. [36:49] Most of them just vaporized. [36:50] This is later, it's five years later, [36:52] during the down cycle in the late 80s, [36:53] and that was around the time that Michael Dell [36:55] in his dorm room decided to get into the PC business, [36:57] and that's exactly right. [36:58] He was a version of that. [37:00] He was a more systematic thinker [37:01] than the Wildcatters who had been [37:02] in the PC industry before that. [37:03] Is that how Oracle wins in databases? [37:05] because there's a ton of database companies back then. [37:08] Yeah, I think Oracle was a somewhat different story. [37:10] I think it might've been more of a story [37:11] of just raw aggression. [37:12] Got it. Larry was always [37:13] very into Japanese Samurai culture, [37:14] and I don't think that was a— [37:16] So moving forward in time, [37:18] why did none of the pre-Google internet companies survive? [37:21] Lycos, Excite, AltaVista, AOL, Yahoo. None of them. [37:25] So, I think that you need to really rewind back [37:28] to the differences between then and now, [37:30] and I would just say a couple things on that. [37:32] One is the whole internet boom bubble, [37:35] whatever you call it, of that period, [37:36] was basically four years. [37:37] It was basically four years in and out. [37:39] For example, the companies you just mentioned, [37:41] for the most part, my company got going in '94, [37:43] those companies really got going in '96. [37:45] By 2000, [37:46] it was nuclear winter, [37:49] and so it was a four-year period. [37:50] The business models either didn't exist or were brand new, [37:54] and we could spend a lot of time on that, [37:55] but all the business models that you have today, [37:57] that have these big mega companies, [37:59] those business models didn't exist. [38:01] It was still mostly just packaged software in those days, [38:04] and so it was really hard to build the kind of [38:05] enduring business that you see today. [38:07] And then I would say the third thing is, [38:08] the market was so small, [38:10] so the total market size in 1999, [38:13] for internet anything was 50 million people [38:15] total, max, maybe. [38:16] Wow. Maybe. [38:17] Half of those people were on dial up, [38:19] which only barely counted. [38:22] By the way, and that was mostly AOL, [38:23] which only barely had internet support [38:25] the way we understand it. [38:26] Yep. Right. [38:27] They had a browser, but it wasn't like what you're used to. [38:30] No. And then the PCs [38:31] were super slow, the modems were slow, [38:33] and that was still, the median internet experience [38:35] in those days was you dial in [38:36] for maybe an hour at night from your desk at home. [38:38] Yep. And then businesses, [38:39] by the way, even businesses that had internet connectivity, [38:42] were doing everything they could to prevent their employees [38:44] from using it. [38:45] How are we doing over here? Everyone all right? [38:46] Guinness? All right? [38:47] All right. Here you go. [38:48] All right. Very good. [38:49] Anyone need anything else? [38:50] Finally, a real Irish— [38:53] Bartender? Finally a [38:54] legitimately-Irish bartender. [38:56] You need a refill? [38:57] That would be fantastic. [38:58] Thank you very much. [38:59] All right. Outstanding. [39:00] So, it was a very early crude time as compared to now. [39:04] So there's another question that leads to, [39:06] which is normally, you get bull and bear cases [39:08] on crypto, or defense, or enterprise SaaS. [39:11] AI seems unique in that there's very little in terms [39:14] of articulate bear cases about why it matters. [39:16] In fact, most of the bear cases go the other way, [39:18] that it's going to destroy the world, [39:19] or something like this. [39:21] Were there articulate bear cases on the internet [39:23] during the bubble? [39:24] Oh, yeah. [39:25] Well, the original bear case was just, [39:26] "Nobody's ever going to make any money. This is ridiculous." [39:29] And then there was just a huge onslaught [39:30] of, "This is just going to be cyber crime, and porn, [39:32] and spam, and fraud, and abuse." [39:34] So you had the similar [39:35] sort of equivalent of AI safety. [39:36] Well, every new technology has a moral panic [39:38] that it's going to ruin society. [39:39] Every consumer technology. [39:40] And then you just use the product [39:42] and be like, "This is a joke. It doesn't really work. [39:44] Look at how long it takes images to load." [39:46] Is anybody really going to [39:47] put their credit card in? [39:49] So, I don't know if bear case is the right term, [39:52] but there was massive skepticism. [39:53] Let's do remember bear case here for a second. [39:55] I think the smartest bear case was that, [39:58] "The internet's clearly a cool thing. [40:00] You guys are getting way over your skis [40:01] in terms of valuations here, [40:03] and in particular, you're getting way over your skis [40:05] in terms of the build out that's happening [40:07] of the internet infrastructure [40:08] where the demand will take awhile to catch up," [40:10] and of course, that was true where there was [40:12] a fiber overbuild, and clearly, [40:16] there isn't an AI bubble in the sense [40:17] that everyone really likes their tokens. [40:20] The stuff that we're doing with AI [40:21] or my personal ChatGPT usage, I really like that. [40:24] You're not going to take that away from me, [40:25] and so it's not a bubble in that regard, [40:27] and it's sensibly priced and everything like that. [40:29] It's a true tech better, faster, cheaper story. [40:32] However, there is a huge ramp up [40:36] in AI data center build out. [40:38] Oracle just had that 4x RPO beat [40:42] that caused their stock to go up 40% [40:44] and Larry Ellison to become the richest man in the world. [40:46] Basically, they're doing giant data center projects [40:49] for AI companies, and one can imagine [40:52] that there will be a data center bubble [40:54] where people get too excited about the build out [40:57] and we build capacity ahead of utilization, [41:00] and people finally, it's the last musical chair, [41:02] people build that data center [41:04] where actually no one wants to lease it. [41:06] Right. Do you think [41:07] that is happening? [41:08] Will happen? Is that a sensible framework? [41:09] So, I would say actually, that is precisely [41:11] what happened with the internet boom. [41:12] Exactly. Right. [41:13] Sorry. That's my analogy. Right, right. That's right. [41:14] And so for people who don't know this, [41:15] what happened with the internet boom [41:16] was there was this sort of internet software [41:18] and services, and Netscape, and Amazon and these things— [41:20] And by dollars, people confuse the dot-com boom. [41:22] Correct. The internet stuff didn't [41:23] matter, it was a— [41:24] It was a Telco bubble. [41:25] It was a Telco bubble. [41:26] It was almost entirely a Telco bubble [41:27] and it was almost entirely a Telco crash, [41:28] and you know that for two reasons. [41:29] One is the sheer amounts of money involved [41:31] were so much greater on the Telco side, [41:32] and then the other is Telco is where the debt came in. [41:33] And to get a really monumental crash, [41:36] depression, recession depression, [41:37] you need a credit bubble to pop, [41:39] and the credit bubble was 100%, I can tell you, [41:42] not on the tech companies. [41:43] It was 100% on the telecom companies, [41:45] and it was massive, and it was amazing. [41:48] And some of them had dodgy stuff going on, like WorldCom. [41:49] Oh, and then there was fraud. [41:50] Right, exactly, and those stories are truly spectacular. [41:53] My retrospective explanation of what happened, [41:55] consistent with what you were saying, basically, [41:56] was, there were a small number of people [41:58] who were building the software and services, [42:00] and that was because they all had to be invented [42:03] from scratch, and then there just only a small number [42:06] of people who even understood the software [42:07] and how you could possibly apply it. [42:09] There just weren't that many of us [42:10] running around who did that, [42:11] and so, John Doerr had a famous saying, [42:12] at some point internet became a cream [42:14] that you rub on investors to get them excited, [42:16] and when that happened, what happened was [42:18] you had a much larger number of people [42:20] who had a lot of knowledge about [42:22] how to put buildings in the ground [42:24] and how to fill those buildings with fiber. [42:26] And the good news with being in the data center, [42:27] in those days it was data centers, right, [42:29] it was data centers and— Hold on, hold on. [42:30] This gets to an interesting point, [42:31] which is— Yes. [42:32] That when you get a boom, [42:34] because the new people, there aren't enough people [42:36] with the new skill set to do it— [42:37] That's right. That can never be [42:38] the epicenter of the bubble. [42:39] That's right. It's always where [42:40] the 50-year-olds with lots of capital are, [42:42] that's where the epicenter is, [42:43] so it was Telco in the internet bubble [42:47] and all those Telco people, sort of 50, [42:49] and so now it's data centers. [42:50] Well, exactly, [42:52] and the way I would describe it is, [42:53] when the thing takes off, [42:55] when the core thing takes off, there's just too much money. [42:57] Yep. There's too much money [42:58] that wants to come in and participate, and it literally [43:00] cannot participate— But also it comes in [43:01] the way it knows how. [43:02] It comes in the way it knows how, [43:03] and this is what you would find at the time, [43:06] which was you would meet a lot, [43:08] and I met a lot of these guys, [43:09] a lot of these were Telco CEOs [43:10] or a lot of these new Telco companies, Global Crossing, [43:13] all these new companies, [43:14] but Global Crossing was one of the great boom, boom, boom [43:17] blow-up stories at the time, [43:18] and the entrepreneur was this guy Gary Winnick, [43:19] and he was actually at Drexel Burnham. [43:21] He was a bond guy from the 80s, a Leonard Brown guy, [43:25] and he figured out like, "Oh, we know how [43:27] to put buildings in the ground. [43:28] We know how to build fiber. [43:29] You go to Cisco, you buy the devices, [43:31] you rig up the fiber, Corning will sell you the fiber, [43:32] and it's a known thing, and his expertise [43:35] was going to the debt market, convince them to finance that. [43:36] And then he could go just like hoover up capital, [43:39] and, in fact, he built like tremendously valuable, [43:41] tremendously important infrastructure. [43:43] It's just that a bunch of that infrastructure [43:44] was not actually filled up for 15 years, [43:47] and in the meantime, much like luxury hotels, [43:50] it traded hands three times. [43:51] The people who own that infrastructure today [43:53] are doing very well with it— [43:54] And so it would be— Many of those companies [43:56] went under. [43:57] It would be ironic if AI researchers are still underpaid, [44:01] but there are too many GPUs per AI researcher. [44:04] Yeah, so, this is the thing, [44:05] and here's where you get into the question [44:07] of whether you can have a reason by analogy, [44:09] and whether things are actually the same, [44:11] and so then it's like, "All right, is AI the new internet?" [44:14] And it's like, "Okay, if AI is the new internet, [44:16] then you could maybe plausibly expect this kind of cycle," [44:19] and for sure, you guys probably meet, [44:21] I meet people all the time, [44:22] which is like, "I don't know how to invest in the software [44:24] side of this, but we're going [44:24] to do a giant data center build," [44:26] and you know, this includes nation states doing this, [44:28] and so you could say history is repeating itself. [44:31] The counter argument to that is, I don't know that AI [44:33] and internet are even remotely comparable. [44:35] Well, another way to say it is, [44:36] if you could have sped up broadband by maybe five years, [44:40] the internet bubble isn't a bubble. [44:42] It just seamlessly goes into 2007. [44:44] You still at 56K modems in 2001. [44:47] Correct. People forget. [44:49] You remember this, but people forget or don't know this, [44:52] home internet broadband was not common until after 2005, [44:55] and I was actually at AOL, I followed this very closely, [44:58] because we sold our company AOL. [44:59] I was at AOL on the executive staff [45:01] in the board meetings in 1999, [45:02] and the big question for AOL at that point [45:04] was how to get the narrow band provider [45:06] to being the broadband provider, [45:08] because we knew it would happen at some point, [45:09] but it was unclear when, [45:11] and ultimately the company couldn't figure it out. [45:12] But the question in those days was very much, [45:15] and it was literally, it was cable modems [45:16] or, it was called ISDN, it was proto broadband [45:19] from the Telcos, and it just wasn't happening, [45:20] and in fact, it didn't happen in that scale in 2005, [45:23] and then mobile broadband didn't really happen [45:25] until 2012. Yeah. [45:28] And people actually forget, the original iPhone [45:29] from 2007 did not have mobile broadband. [45:31] Or apps. [45:32] Or apps. Right. [45:33] But it also, it was on the old AT&T— [45:36] It was useless. [45:37] Yeah, yeah, yeah. Yeah [45:38] Yeah, so there was this just incredible lag [45:40] for when an ordinary person could have [45:41] the kind of experience that you can have today. [45:44] And so, yeah, so one theory [45:45] for why, quote, AI is different, [45:47] is actually, no, the experience that you're having today, [45:49] just in ChatGPT, it's just so monumentally amazing. [45:52] It's fully there, and yeah, you have to watch it [45:54] type the thing out, but the answers are like spectacular. [45:57] Yep. And so there's that, [45:58] and then there's the other thing, [45:59] which is just the problem with metaphors, [46:02] which is, one of the theories you could say on this is, [46:05] the internet was a network technology, [46:08] whereas AI is a computing technology, [46:09] and maybe the only comp comp for AI [46:12] that you can have is actually the creation of the computer, [46:15] because it's the first major reinvention [46:17] of the fundamental model of what is a computer [46:19] in 80 years, going from the Von Neumann architecture [46:21] to the neural network. [46:22] And if you trace the history back, [46:23] they knew in the 1940s that these were the two paths. [46:26] Yeah. [46:27] They knew what the neural network was in 1943. [46:29] There was a big argument at the time [46:30] of whether the computer should be based on [46:32] fundamentally adding machines on cash registers [46:35] or whether it should be based on brain architectures, [46:38] and we had to wait 80 years for it to work, [46:40] but now we have the computer industry V2, [46:42] which is much more valuable and important [46:44] because of all of the obvious things it can do [46:46] that the hyper-literal Von Neumann machines can't do, [46:49] and so, we've successfully unlocked computer industry V2, [46:52] it's 10, or a hundred, or a thousand, [46:54] or a million times more important and valuable, [46:56] and all of your petty comparisons to [46:59] bubbles in the 1990s just wash out, because, my God, [47:02] look at what the thing could do. [47:04] AI is funny, because it is always the case [47:06] that the hype cycle for technologies predates the technology [47:10] being ready for that hype, [47:11] and so, Charlie and I often talk about [47:13] the mobile internet hype. [47:14] Yeah, that people were excited about. [47:17] You'll buy cinema tickets on your mobile phone [47:18] in the 2000s on a Nokia 3310, which is not actually [47:21] how the mobile internet played out, [47:23] and even the crypto excitement, [47:25] the kinds of things people talk about with crypto, [47:27] of like, "Oh, you'll be able to make payments," or whatever, [47:29] we're finally getting to it in 2025 [47:31] in any kind of meaningful volumes, [47:32] but it took a good 15 years [47:35] from when people started being excited about it. [47:37] AI is maybe the longest time lag [47:39] from those things where, [47:42] when was "2001 A Space Odyssey" released? [47:44] The books that it was based on, were in the 1950s, [47:46] and then "2001 A Space Odyssey" [47:48] was the 60s? 60s. '68. [47:49] Yeah, yeah, yeah, exactly. '69, yeah. [47:51] And that was voice mode with tool use, [47:54] like HAL 9000, [47:56] and so, I find it funny that we had such a specific vision [47:59] that was pretty much right, [48:02] but it took a long time for the tech to be... [48:04] And there was various waves, dragon systems, [48:07] the tech wasn't that good, [48:08] but people were excited about it. [48:11] There's a book called "Rise of the Machines" [48:12] that has the prehistory of AI, [48:13] and, I believe I remember correctly, [48:15] there were actually debates about this in the 1930s. [48:17] It actually predated even the invention [48:19] of the neural network— [48:20] Okay, so roughly a hundred years later, [48:21] we're getting around to it. Yeah, yeah, [48:23] and I think Alan Turing and folks like that [48:24] were involved in that at that time. [48:25] There was a famous moment in the history on this. [48:27] So, Alan Turing, Claude Shannon, [48:30] the inventor of information theory, two very important guys, [48:32] during World War II, they're building the computer [48:34] originally in World War II, to beat the Nazis, [48:36] crack the codes, and so Alan Turing [48:37] and Claude Shannon are having lunch [48:39] at the AT&T executive dining room [48:41] in Basking Ridge, New Jersey, in 1943, [48:44] and they're talking about exactly this topic, [48:46] and Alan Turing starts to raise his voice, raise his voice, [48:48] and finally, he gets up in the middle [48:49] of the AT&T dining room and says, "I'm not talking about [48:52] building a genius computer brain. [48:53] I'm talking about building a mediocre computer brain, [48:55] like the president of AT&T." [48:57] Right? [48:59] And so, they knew, I think he knew that the path [49:02] that they were on, the Von Neumann-machine path [49:05] was building this hyper-literal, [49:08] you can almost say hyper-autistic math savant in a box, [49:11] which obviously, was not going to be the thing [49:14] that was going to be English language [49:15] and write everything else that you were going to want to do, [49:17] and so he knew this is the wrong path, [49:20] but he just didn't live in the time in which the technology [49:22] was available to do what he wanted to do, [49:23] and it just happens that we do. [49:24] What are these things emerging heuristics [49:27] of how the market works? [49:28] So, let me give an example from software. [49:29] There's no inferior-goods market for software. [49:32] There's no cheap version of Excel, [49:35] there's sort of one— There was at one point. [49:36] There was at one point, [49:37] and it didn't succeed, which is the point— [49:38] They'll throw it out. But in general, [49:39] software's gone to one company, some horizontal, [49:42] some vertical, being the best— [49:43] because they're such a great deal. [49:44] because it's such a great deal. [49:46] A hundred percent. A hundred percent. [49:47] Is it the same in AI? [49:48] Do we go with horizontal intelligence? [49:51] Is there an inferior-goods market where [49:54] you end up with AI and device, it's intelligent [49:56] but not super intelligent, [49:57] but you don't need it— Would people use— [49:58] To ask the weather? Worse models? [49:59] The way I would think about it is, if you think about, [50:02] let's say this is a computer industry V2, [50:04] what did you experience in computer industry V2? [50:06] You had many different sizes and shapes of computers, [50:10] and actually what happened [50:10] at the time was the big ones got built first, [50:12] and then it literally was mainframe, [50:14] and then it was mini computer, [50:15] and then it was server, [50:17] and then it was— And then it goes down [50:19] to the PC. Personal computer, [50:20] and then mobile phone, and then embedded devices. [50:22] Yep. Right? [50:23] And then, by the way, and then it multiplies outward. [50:24] Cars, and light bulbs, and doorknobs, [50:26] and everything else. [50:28] As you know, what you have as a consequence, [50:29] is the computer industry, [50:30] and specifically, the chip industry, [50:31] is therefore in the form of a giant pyramid, [50:34] where at the top, you have a small number [50:35] of supercomputers and mainframes, [50:36] and at the bottom you have billions [50:38] and billions of embedded devices, [50:38] and then you have everything else in the middle. [50:40] And the reason you have that is because [50:42] you have custom performance and fit implications [50:45] for the specific devices. [50:46] You don't want your light bulb [50:48] to have to do a round trip to an IBM mainframe or something. [50:52] It doesn't make any sense. [50:53] You want it to have the embedded device [50:54] so that it senses whatever you want. [50:56] Senses whether there's light in the room. [50:58] That's just a specific chip, [51:00] and so, I think the scenario in which you only have [51:02] a few big AI models is the scenario in which [51:04] not only are those models the smartest, [51:06] but they're also the cheapest, and the most power efficient, [51:09] and the fastest, and easiest to adopt and use [51:13] for every scenario, [51:15] and I think that's highly unlikely [51:16] just because, if this is the breakthrough [51:18] that we believe it to be, and it's the computer industry V2, [51:20] you're going to want models in everything. [51:23] So you're going to want AI infused into everything, [51:26] and then, for a lot of those infusions, [51:28] you don't need your doorknob to teach you quantum physics, [51:31] but you do need it to be really good [51:32] at knowing that it's you and not somebody else. [51:34] Yep. Right? [51:34] And so you're going to have all of these [51:36] hyper-optimized use cases, [51:37] and so, my guess in the way we're betting [51:40] is you're— Everywhere all at once. [51:41] You're going to have that pyramid approach. [51:42] Yeah. And then, look, [51:44] the economics are going to be a big part of that, [51:47] if only because the doorknob gets to run a local power, [51:50] and then the processing the doorknob needs to do [51:53] is a tiny fraction of what you need to do [51:54] when you ask GPT-5 a query. [51:56] And so, I think this is computer industry v2 in that way. [51:59] And how do the markets play out where, [52:00] is it just a normal battle price performance [52:03] with proprietary players? [52:05] How big a player is open source here? [52:08] Charlie mentioned Oracle earlier. [52:10] I feel like people today forget [52:11] that the proprietary databases [52:13] used to be the best databases [52:15] all the way through the 90s, [52:16] and you had to, step one of founding an internet company, [52:18] was write a check to Oracle, [52:20] and then you can do stuff after that. [52:22] And then the open source databases, [52:25] most people in post-grads became competitive [52:26] in the 2000s. [52:29] You don't like me reasoning by analogy [52:30] too much here, [52:31] but can you reason by analogy to the database world? [52:33] Just how does the market structure play out? [52:34] Yeah, no, I think that's right. I think that's right. [52:36] I think that's actually a good comp. [52:37] Another one is operating systems. [52:39] So, when I was a kid, [52:41] the world's best operating systems were specifically, [52:44] Windows is its own trajectory, and iOS, [52:46] but for what we used to describe [52:49] as proper computing on real computers, [52:52] like Unix computers, including supercomputers [52:55] and workstations, and advanced scientific applications, [52:58] things like that, the best versions of Unix [53:00] were proprietary for a very long time, [53:02] and you had these really big companies, like DEC, [53:04] and HP and others, IBM, [53:06] that had their own versions of Unix [53:07] and they made a lot of money on those, [53:09] and then Linux, same story. [53:13] Linux came along, looked like a toy, [53:15] and then 10 years later it was better [53:16] than all the proprietary ones [53:17] and all the proprietary ones died. [53:19] That's my guess, is that something like that. [53:22] I definitely think we'll live in a world [53:23] of a small number of big models [53:25] that'll be incredibly valuable [53:26] and incredibly widely used for many things. [53:28] My guess is we are going to live in a world [53:30] in which most aggregate AI is going to be executed [53:32] probably on smaller-form factors, [53:35] and probably most of that is going to be open source. [53:37] So where is ground zero, where the rate of change [53:38] would be highest? [53:40] Software development? Somewhere else? [53:42] Software development [53:43] is a very good candidate for that, [53:44] just because you have people building [53:45] for themselves, I think, [53:47] and you have this incredibly tight iterative loop, [53:50] and you see that with these AI tool companies, [53:53] so that's the claim, [53:55] and then, by the way, the other advantage [53:55] of software development is, [53:56] this is a really underrated thing [53:58] with respect to AI adoption [53:59] that a lot of the people in the field are missing, [54:00] is software development is not regulated, [54:02] and so it's like impossible— Yet. [54:05] Well, there is that. They are trying. [54:08] The enemies of progress and freedom are trying, [54:10] and we are fighting them very hard, [54:12] but AI medicine actually can't move that fast, [54:14] because it's regulated, and AI can't be a doctor. [54:16] It can't get licensed. [54:18] AI can't be a lawyer, [54:18] it can't go make an argument at a court, [54:21] and so forth and so on, [54:22] and so, the unregulated fields populated by [54:27] the same kinds of people who are building AI. [54:29] Charlie had the interesting question [54:30] of are we overestimating the broad impact [54:34] and underestimating the specific impact, [54:36] or what of, at least for the next five years, [54:38] as you say, AI in medicine or AI in law [54:40] doesn't make that much progress [54:42] because of some of the challenges, [54:44] but software engineering is totally transformed. [54:46] So the counter argument, I think there's a big argument [54:48] in that direction, and, by the way, I actually wrote [54:51] a big Substack piece maybe we can link to, [54:52] talking about how the employment shifts [54:54] everybody's worried about are actually not going to happen [54:55] at anywhere near the velocity people think, [54:57] because a significant percentage of jobs [55:01] in the US literally are licensed or unionized, [55:03] or civil service in a way where they literally [55:05] cannot be replaced, [55:06] and so, I do think there is part of that. [55:08] Having said that, I think things are going to pop [55:10] in really interesting ways, [55:10] and so, for example, ChatGPT is in fact [55:13] a better doctor than your doctor today, [55:15] with almost a hundred percent certainty, [55:16] and just the fact that it can't literally be your doctor [55:19] doesn't mean you're not going to ask all the [55:20] doctor questions. [55:21] And then you already have people online [55:22] who are taking surreptitious camera-phone footage [55:24] of their own doctor asking ChatGPT during the appointment. [55:28] I think the medicine use case is an interesting one, [55:29] because it turns out it was a space [55:31] where most people were actually intelligence bottlenecked, [55:34] which, it's like test-time compute, [55:36] they were getting a very small fraction [55:38] of their doctor's headspace, [55:40] and if you put just more thought on the problems, [55:42] you can get really good outcomes. [55:44] And, by the way, medicine and law, [55:46] you could also look at the self-driving car thing, [55:48] which is, there's always this test for self-driving cars, [55:51] there's always been this question of [55:52] is the requirement perfection [55:54] or is the requirement better than the median human driver? [55:57] And if you apply that same question [55:58] into law or medicine, it's just overwhelmingly clear [56:01] that you're better off today with doctor ChatGPT. [56:03] Now, in one sense, you can't live your life that way, [56:06] because it can't be your doctor. [56:06] On the other hand, you can sit there all day long [56:08] talking to it about your health. [56:09] Yep. And, by the way, [56:10] I think there's going to be a lot of tension [56:11] and a lot of drama in these different fields [56:14] as that happens, but here's another ar argument [56:16] that comes back around on this, [56:17] which is the argument of "Oh, AI's horrible [56:19] because it's going to lead to five companies [56:20] controlling everything," and that's it. [56:24] The monopoly cartel fear, [56:26] and there's a bunch of reasons to be suspicious of that, [56:28] including things like open source, [56:29] but the other reason to be suspicious is, [56:31] at least with downstream impact is AI is already [56:34] maybe the most democratically-distributed [56:36] technology in history, [56:38] so 600 million people or whatever the number now is, [56:41] on ChatGPT in two years, [56:43] and again, you compare that to internet adoption, [56:45] it's far faster, and of course, [56:47] the reason is because the internet exists today [56:49] to be able to distribute it, [56:49] but the world's most advanced AI is in an app [56:53] that 600 million people have. [56:54] It's not in the one that I have or that you have. [56:56] It's the one that 600 million people have, [56:58] and so this technology has already been hyper-democratized, [57:05] and so it's going to be in everybody's hands. [57:06] And people get confused about this, [57:08] because they're like, "Well, why would [57:09] big companies do that?" [57:10] And the reason is because the mass market's [57:11] always the biggest market. [57:13] You want to get to everybody if you're trying to build [57:15] the most successful company, [57:16] and to be the company that is the most important, [57:18] and look, for sure there are always concerns [57:20] about aggregation of power [57:21] and centralization of power, for sure, [57:23] but there's this other thing which is, [57:25] what if this is just the philosopher's stone, [57:27] the alchemy of sand into thought, [57:30] in literally everybody's hand, [57:32] right out of the gate? Well, so there's a... [57:34] if you look back at the old companies, [57:36] you look at the S&P 500 of some 1980, [57:39] there's not that much change in the success based on tech. [57:43] I.e. it's not like some bank gets better tech [57:45] than all the others and just goes past all the competitors. [57:48] If what you're saying is true, [57:49] you would say that old companies [57:50] are going to adapt less well [57:53] to this. Oh, 100%. [57:54] And the level of change is going to be unprecedented. [57:56] Yeah. I believe that to be the case. [57:57] Well, so again, let's go back [57:58] to the computer industry on this. [57:59] This, I think is a very interesting idea. [58:01] So, the computer industry started out [58:03] by building the big thing. [58:05] Started out by building the mainframe. [58:06] Thomas Watson Senior, who ran IBM in the 1950s, [58:08] said he thought there was a world market for five computers, [58:10] and it was literally one mainframe each [58:13] for the three big insurance companies, [58:14] and then two for the Department of Defense, [58:15] and that was it. [58:16] And, by the way, at that time, it was true. [58:18] That was the world market— [58:19] For those computers. [58:20] For those computers at that time, [58:21] and then basically, over 40 years, he went from mainframe, [58:24] to mini computer, to client server, to, as you said, [58:26] to PC, and to phone, and so what happened is over 40 years, [58:28] the technology cascaded down into the mass market, [58:32] and then today, it's culminated in [58:33] the $10 Android smartphone in India, [58:37] and so, that was that. [58:40] AI, at least so far, and, by the way, [58:42] many other categories of new technology [58:43] in the last 30 years, because smartphone [58:45] is another example of this, [58:46] or have been the reverse, which is no, [58:48] the individual gets it first. [58:50] The companies are deciding to go [58:52] for the individual market first, [58:53] because that's the largest market, [58:55] and those are the people who are [58:56] the easiest to adopt. It's Andy Warhol, [58:57] "The president drinks the same Coke as you and I." Yeah. [58:59] Exactly, and then what happens is, over time, [59:03] and this is what I think happens with smartphones, [59:05] and this is what I believe is happening with AI, [59:07] which is the individuals get it first, adopt it first, [59:09] the small businesses get it second, adopt it second, [59:11] the big businesses get it third, [59:13] and the government gets it forth. [59:14] Not because the governments and the big companies [59:16] couldn't get it faster if they wanted to, [59:18] but they can't, because they can't absorb it. [59:21] They have all their rules [59:22] and then have all their bureaucracy [59:24] and they just simply can't absorb it, [59:26] and so, and again, at the level of politics, [59:29] structured society, you could say this is like [59:31] a fight between the power of the individual [59:33] versus the power of the state. [59:35] Obviously, there's fears of like AI surveillance [59:37] and all these things on the state, [59:38] but the other side is every individual citizen [59:40] being super empowered, being a PhD in everything, [59:42] including how to deal with the state. [59:44] Right? [59:45] So everybody, all of a sudden is a super lawyer. [59:47] Yep. Okay, and then, [59:48] within business, it's the balance of power [59:50] between small companies and big companies. [59:52] And, if you're just looking at speed of adoption, [59:54] there's no question small companies are adopting faster. [59:56] I was going to ask about that, because Robert Solow said [59:59] the computer age shows up everywhere [60:00] except the productivity statistics, [60:02] AI productivity is showing up everywhere [60:04] except the hiring plans of your portfolio companies, [60:06] which still seem to be hiring a lot of humans. [60:09] What does the realization of significant [60:11] AI productivity-gains look like? [60:13] Because presumably like stodgy, large companies [60:15] you believe will fight the gains at some level. [60:18] They won't take as much AI productivity as they should. [60:20] So, I think the most basic question, [60:22] the fundamental question of, is this a centralizing power [60:25] or is this a democratization of power? [60:27] And— So do you think [60:28] it'll make smaller companies more powerful [60:29] in the battle against large companies? [60:30] I think there's a really good chance of that. [60:32] I don't know for sure and we'll see. [60:33] But it seems certain that it will make younger companies [60:36] more successful against older companies. [60:37] I would assume so, or the kinds of— [60:40] Less bureaucratic. That have the kinds of... [60:42] Right. Exactly. [60:43] Less bureaucratic, but let's take the employment [60:47] and the jobs thing, just because that's the one [60:48] that gets all the headlines, [60:49] which is it's just like, "Oh, all the jobs are going to [60:51] go away, because AI's going to do everything. [60:52] So one version of it is, okay, that is going to be the thing [60:55] of it, and this leads to the meme [60:57] of five companies are going to own the world [60:58] and you have, whatever, three years to get out [60:59] from the permanent underclass. [61:00] Yeah. Right? [61:01] Whatever, whatever. Right, this leads to that. [61:03] The more conventional economic argument [61:04] is the opposite argument, which is, this is going to deliver [61:07] massive productivity improvements, [61:10] not just to companies, but also to individuals. [61:12] When you put a technology in the hands of an individual, [61:15] it massively increases their productivity, [61:17] and the way I think about that is AI [61:19] just makes every individual a super PhD in every topic. [61:21] That's the most dramatic increase [61:23] in what economists call marginal productivity of the worker [61:25] that has ever existed, [61:27] and so, as a consequence, [61:28] every single one of those people is now capable of doing [61:30] so much more than they were ever capable of doing before. [61:32] Whether they're doing that as a solo entrepreneur [61:34] or whether they're doing that as somebody [61:35] who works in an organization. [61:37] And so, in that version of the world, [61:39] you don't get the aggregating effects. [61:41] You get some, but they're swamped by the democratization [61:43] and the superpowers that every individual gets, [61:45] and then 10 years from now, [61:47] we'll do part two of this, [61:48] probably with the same glass of beer, [61:50] with the same room temperature, [61:53] and we will be shocked by how much AI drove [61:55] both employment growth and drove incomes, [61:59] because again, the conventional economic view is [62:01] marginal productivity improvements, [62:02] you want to hire more people at higher [62:03] levels of productivity, [62:05] because they can do more, and then you pay them a lot more [62:07] because they can help at a higher rate- [62:08] A huge part of that is, when people think about this, [62:10] is intelligence but not imagination. [62:11] Right. If you go back to 1950, [62:13] there's some movie there where basically a single person [62:17] is a cell in Excel. [62:18] They're all sitting in a big room, effectively, [62:21] doing accounting. [62:23] If you described the computing evolution, [62:26] they would all say, "I'm going to lose my job," [62:27] but the jobs that emerge, video gaming, [62:30] you couldn't imagine, you couldn't describe, [62:32] so it's very hard, I think, for people to overcome [62:35] the jobs they can see existing disappearing, [62:37] but they can't see the emergence of new categories— [62:38] That's right. But we've always had [62:39] the emergence of those new categories. [62:41] That's right. And if you take things [62:42] like sport, which I think is like 3%, 4% of GDP, [62:45] you can imagine that extending to 20% of GDP, [62:48] and whole new sports emerging. [62:51] If you get more GDP— We already have [62:52] whole new sports emerging with eSports, [62:53] and, so you can argue all sports have gotten way bigger [62:58] over the past five years. [62:59] Like basketball is way bigger, F1 is obviously way bigger. [63:01] Just, they've all gotten much bigger. [63:03] Yeah. We're even bringing soccer [63:04] to the US, which— [63:05] Exactly. [63:06] It's even inconceivable. [63:07] No, that's exactly right, [63:08] and yeah, and then the corellator to that, [63:10] by the way, this is very difficult to talk about, [63:12] because people get very upset, [63:13] but the correlator to that is those old jobs, [63:15] after the fact, you're just like, "I can't believe [63:16] human beings were required to do that," [63:19] because literally, like as you're alluding to, [63:20] what happened, [63:21] the original— Back-breaking Excel work. [63:23] The original computer [63:25] was a person sitting at a desk doing manual math [63:27] all day long. [63:28] Imagine if I showed up today and told you [63:29] that's what your kids are going to be doing as a profession. [63:31] You'd be like, "Sounds like torture." [63:32] Have you read ever, Iain M. Banks, [63:34] the science fiction author? [63:36] "The Culture Suits." [63:37] No, I've actually never read that. No. [63:38] Okay, he tries hard to contemplate [63:41] what a super-advanced society with AI is like, [63:43] and what's interesting is, everyone has stuff [63:45] that looks like a job, but is actually leisure. [63:49] Right, well, the best jobs in the world have [63:50] that characteristic, right? Yeah. [63:52] And then you have very complex status hierarchies [63:54] as people aspire, and if you look [63:55] at Gemeinschaftic societies, like Formula One [63:58] or something like that, [63:59] you have a very clear motivation [64:02] states hierarchy for people within it [64:04] that seems to fill [64:05] a lot of human needs. Aren't you describing [64:06] being a VC? [64:07] I've seen the activities [64:08] that lots of VCs go to. Yes! [64:09] Exactly, as we like to say, [64:11] it's a nine-through-four profession, [64:12] so— Exactly. [64:14] It's a country club kind of thing. [64:16] The other, by the way, great economic fallacy [64:17] that I just see everywhere right now, [64:18] is this idea that AI is somehow going to be [64:20] this hyper-successful thing. [64:22] Hyper acceleration of productivity [64:23] and dramatically change everything, destroy all the jobs, [64:25] and yet somehow that's going to lead [64:26] to people being eviscerated, and being poor [64:28] and not having anything, [64:29] and the missing element there is that [64:31] even if that scenario plays out, which I think, [64:33] as I said, I don't think it's a centralization scenario, [64:35] but even if it played out, [64:36] the result would be hyper deflation of prices, [64:39] which is the thing that people miss, [64:40] and so, in that environment, [64:43] with that level of productivity growth, [64:44] the price of business services will collapse, [64:47] and things that today cost a lot of money [64:48] will all of a sudden all be cheap or free. [64:50] This is sort of the— [64:51] Everything becomes oversupplied. [64:52] In Star Trek, GDP would be zero. [64:55] Right. Because you just— [64:56] The replicator does everything. [64:57] The replicator does everything. [64:58] Right, and so things that cost [65:00] a hundred dollars cost a penny. [65:01] In that world, even real GDP looks like it's shrunk, [65:06] and everybody is much, much, much, much better off, [65:09] and by the way, this is not the first time this is, [65:09] there have been periods of sustained deflation in the past. [65:12] Well, you see it within categories. [65:13] Look at the spend on CDs. [65:15] Yeah. Yeah. Music CDs versus [65:17] music today. So, I always talk [65:18] a lot of the so-called second industrial revolution, [65:21] so the time in which our entire modern world was built [65:24] with everything from airplanes [65:25] to freeways and everything else, 1880 to late 1930, [65:28] it's that 50-year stretch, [65:30] and for a lot of that period, [65:32] they were in essentially a protracted [65:33] deflationary depression, [65:35] because what happened was the technology [65:37] for acquiring and processing raw materials [65:39] was advancing so fast that there were gluts in all [65:42] the different raw materials, [65:43] and so it felt like the economy was caving in, [65:46] because prices were collapsing, [65:47] economic activity was down, GDP was down. [65:50] In reality, what happened was a massive surge [65:52] in productivity growth and a massive surge [65:53] in material prosperity, [65:55] and over that period, both productivity growth [65:56] and economic growth advanced something like 3x of our time, [66:02] but if you read the books at the time, [66:03] they're obsessed with this problem of like, [66:04] "Oh, my God, there's this oversupply of iron, [66:06] what are we ever possibly going to do with it?" [66:08] And it's destroying the economics [66:09] of the iron-production business. [66:11] Could you not have low-productivity segments [66:16] of the economy find ways to [66:21] avoid the prices collapsing too much— [66:25] Yes. Such that you [66:26] don't get this effect, and people— [66:28] Baumol's cost disease. [66:29] Yeah, we've gotten much better at healthcare [66:31] over the past 50 years, and yet. [66:34] Yes, yes, so, Baumol's cost disease, [66:36] but also just simply, government. [66:37] You see this today. [66:38] So, basically, it today what happens if you chart, [66:41] basically, the prices of products [66:43] across all these sectors, what you see is— [66:45] The deflationary economy and the inflationary economy. [66:46] Yeah, there's two different economies, [66:48] and the deflationary economy is everything electronic, [66:50] everything software, everything media, [66:52] by the way, basically everything, all light manufacturing, [66:54] clothes and everything. [66:55] Well, not housing. [66:57] So, the price of clothes collapses, [66:58] the price of housing hyper-rises— [67:00] Oh, sorry. Yeah, they rise. They deviate. [67:02] On the other side, on the non-productive side, [67:04] you've got housing, education, and healthcare, [67:06] and that, I think, explains a lot of the politics [67:10] and feeling of our society right now, [67:12] which is just, everything that's optional [67:15] and fun is getting super cheap, [67:16] and everything that's actually necessary [67:17] to raise a family is getting like hyper-expensive, [67:20] and, exactly to your point, [67:21] it's because these are two different economies, [67:23] and then you look at, and this gets complicated, [67:26] but if you look at housing, and healthcare, [67:27] and education, what they all have in common is [67:29] heavy government interference, [67:31] specifically of the form of restricting supply. [67:33] In all cases, the government basically restricts [67:35] how many houses can get built, [67:37] they restrict how many doctors can get licensed, [67:38] they restrict how many universities can get accredited, [67:42] and then because restricted supply [67:45] leads to prices skyrocketing, the voters get mad [67:47] and so then, the politicians subsidize. [67:49] In all three of those markets, [67:50] there's massive government subsidies, [67:52] federal student loan programs, [67:53] federal mortgage programs, federal healthcare programs, [67:55] and, basic economics, [67:57] if you constrain supply, you cause prices to rise, [68:00] and if you subsidize demand, you cause prices to rise. [68:04] I think this is basically the state [68:05] of the Western democracies over the last 50 years is, [68:08] every step of the way, as the price of what they call [68:12] the American dream, housing, education, and healthcare, [68:17] as the price of those rise, [68:18] the pressure from the government to subsidize increases, [68:22] which just drives the prices higher, [68:23] and so, you're in this ever-escalating spiral. [68:25] I'm presuming you're concerned about more of that. [68:27] Everyone's making fun of the Boston City Council, [68:30] objecting to Waymo, and maybe voting [68:32] to preserve driving jobs and everything like that, [68:34] and so we find more categories [68:36] to turn into healthcare or education. [68:38] Yeah, so sinecures fundamentally [68:39] and then Baumol's cost disease kicks in, [68:41] because then you have the hyper [68:45] incomes being earned by people [68:47] in the deflating sectors [68:48] where there's massive productivity growth, [68:50] and then people in healthcare [68:51] get to command those wages, [68:52] and the whole thing compounds [68:53] and gets worse. [68:54] By default, this is what the governments are going to do. [68:56] In fact, by default, it's what they're [68:59] exactly doing today, [69:01] and then there's a really tricky political-economy [69:03] thing to this, which is, [69:04] it's very hard to tell the voters, [69:06] "Don't vote for the guy who says [69:07] he's going to subsidize housing more." [69:08] Right. Right? [69:09] So, are you worried about this as a political future? [69:12] Yes, 100%. [69:12] Well, I think this is our political present. [69:17] Sure, sure. I was thinking an expanded version. [69:18] Oh, an expanded version, 100%. [69:20] I'll give you the latest example of this. [69:21] Latest example, so remember the dock workers strike? [69:24] Oh, I do. Okay. [69:25] Remember the guy with the gold chain, the whole thing, [69:27] and we found out about the dock workers [69:29] and you dig into it, and you're like, [69:30] "Oh, the dock workers union, it's the reason [69:32] why we've got—" Reason why European ports [69:34] are way more productive than the US. [69:35] Automated ports, because you have unions, [69:38] and they have a tremendous amount of political stroke. [69:39] One of the things that was discovered [69:41] during that process that I didn't know, [69:42] is that in prior union agreements with the dock workers, [69:45] they already had a one-to-one ratio [69:46] of people sitting at home doing nothing [69:48] to every productive dock worker as a consequence [69:50] of the last, whatever, 60 years of these things. [69:52] So, basically, there's a long history here [69:54] that just never became visible in public, [69:56] which is every time any kind of new automation [69:58] shows up at the docks, [69:58] the dock workers renegotiate the contract [70:00] to preserve the jobs, [70:01] which literally means people sitting at home, [70:03] and that was before the most recent agreements, [70:06] and that's just a micro example [70:07] that's easy to pick on. [70:08] The much larger example is the civil service, [70:11] public-sector unions, obviously, [70:13] and then here we're into teachers unions [70:15] and nursing unions and— [70:16] Yes, yes. All of these things. [70:18] And then, here we're into this fairly amazing [70:21] bizarre world we've been in for the last 50 years, [70:23] where you have, especially around government, [70:24] you have both civil service protections [70:26] and union protections, so, exactly. [70:29] So, by default, the political economy [70:31] makes all of this worse and worse. [70:32] By the way, this is why I think inflation [70:35] doesn't mean what it used to. [70:36] Inflation 50 or 100-years-ago used to mean [70:38] the price of raw materials was so important [70:40] in the economy that you felt it very directly. [70:43] Now, that, as you say, you've got this— [70:45] It's hard to talk about a single bundle. [70:46] Yeah, because it's not the same thing, [70:48] and this is the thing, where you can't build a family [70:51] off of the price of the iPhone. [70:53] Just because everybody has infinite media on their iPhone [70:55] for free does not mean that they feel good [70:57] if they can't buy a house. [70:58] I liked your, what was your inflation stat? [71:02] Of if there's a hole in your drywall, it's cheaper to put [71:06] a flat-screen TV over it [71:07] than it is to repair the drywall. [71:09] 100%. Exactly. [71:10] Let me drag us back to AI a sec. [71:13] You used to have the 10x engineer, [71:16] you going to have the 1,000x engineer with AI? [71:17] Yeah, for sure. I think you already do, in practice. [71:20] Okay. And, of course, [71:21] we have had for a long time. [71:22] I mean we've had the 1,000x engineer for a long time, [71:24] it's just, it's becoming— We're going to add [71:25] another order. Yeah, yeah, yeah. [71:26] It's becoming more visible. [71:27] It's going to apply in more areas of software, [71:29] and then, look, the other thing is [71:30] just the payoff to software has been rising, [71:32] because the markets are so much larger now. [71:34] This goes back to the why would this time be different [71:35] with AI versus the internet? [71:36] Which is just, well, okay, this is the first time [71:38] in human history that you've had five billion [71:39] people connected on an interactive network, [71:41] and if you are a provider of products [71:43] and services that go into that market, [71:45] if it worked, it may or not work, [71:47] but if it works, it can get infinitely large, [71:49] and actually really fast now, [71:51] and so, what is the upside? [71:54] How many people are there in the world [71:58] who are going to pay whatever it is, [71:59] 20 bucks a month for the world's best AI? [72:01] So, you can— It's not all five billion, [72:03] but it's a much larger number [72:05] than you would've had 10, [72:07] or 20, or 30 years ago, [72:08] and maybe it's just simply market size. [72:14] As you just heard from Marc, we're in the midst [72:16] of a massive platform shift with AI. [72:18] It's like the computer industry v2, [72:20] and Stripe is the company building the economic [72:22] infrastructure for AI. [72:24] If you've used an AI product recently, [72:26] you've almost certainly used Stripe. [72:27] More than three quarters of the Forbes AI 50, [72:29] including OpenAI, Anthropic, Lovable, ElevenLabs, [72:32] Perplexity, Cursor, Midjourney, [72:33] they all use Stripe to monetize their products, [72:36] and they're growing at a historic pace. [72:38] We analyzed the top hundred AI startups on Stripe [72:41] and found they reached the million dollar revenue milestone [72:43] four months faster than the SaaS companies [72:45] that preceded them. [72:47] AI companies choose Stripe for high converting checkout, [72:49] instant global reach, and AI-driven payments performance, [72:52] all while monetizing in the ways that they need [72:54] with subscriptions and usage-based billing. [72:56] And then as AI agents begin to buy on our behalves, [72:59] Stripe is building the tooling for agent e-commerce [73:01] so that any business can thrive in this next era. [73:04] So come to Stripe to explore our MCP toolkit, [73:07] agentic payment flows, [73:08] and everything your AI business needs. [73:12] You've been super early to crypto with a16z. [73:15] It's probably the area where there's been the most [73:18] concentration of VC performance. [73:21] You, Paradigm, not that many others. [73:24] Two questions. Why did so few VCs focus on crypto? [73:27] Yeah. And then how important [73:28] is stablecoins going to be? [73:30] So, sorry, the first question is why did they, [73:31] or why didn't— Why didn't they? [73:32] Didn't they. [73:33] So, what I've— Status or...? [73:37] So one is you could always just say the easy explanation [73:39] is just they didn't understand it [73:40] or they were focused on other things. [73:41] What I've observed is that, [73:42] as technology has become more important, [73:44] people's belief systems have a lot more [73:45] to do with technology, [73:47] so your worldview, [73:49] the part of your brain that thinks about things, [73:51] a larger and larger percentage of that [73:53] is devoted to technology, [73:54] and, of course, if you're a VC, [73:55] that's a hundred percent, [73:55] and then whatever you're spending your time on, [73:57] you form whatever myths, legends, religion, cults, [74:04] so it means the same question [74:05] of why is the press so much more focused on technology [74:07] than they were 15 years ago? It's— [74:09] Essentially, it was harder to be politicized [74:11] about AOL and eBay— [74:12] Yeah, who could get— Than it's today. [74:13] Yeah. Exactly. [74:14] I think what we observed is, a lot of VCs [74:16] who were very logical and passionate [74:18] on topics like SaaS [74:21] for which there's no religion. [74:22] It's hard to get political [74:23] about SaaS. It's hard to get political [74:24] about SaaS, and for some reason, [74:25] there was something about crypto [74:26] where they just got locked in on, yeah, [74:28] the politics or the whatever, [74:30] and it was just like... [74:32] Oh, so my theory of it after a while, [74:34] because I just met so many people [74:35] who would just flow with it, [74:36] and it wasn't even that they were like, "Ah, we don't think [74:37] it's going to be valuable." [74:38] They were going to be like, "Oh, it's evil." [74:39] "It's full-on evil." "It's a scam." [74:41] "It's a fraud." "It's a this." [74:42] "It's a that." [74:43] If it works, it's evil. If it doesn't work, it's evil. [74:46] One of one my tentative conclusions, [74:48] was just money pisses people off, [74:50] and so, making money through tech [74:52] is usually an indirect process. [74:53] In this case there was a [74:54] more direct aspect— I see that. [74:55] It interesting. People, really, [74:56] people have always built up all kinds of weird religious [74:58] and political views around money. [75:00] And yeah, so what we experienced [75:02] was people just got really upset, [75:04] and we could never understand it, [75:05] because what's the point of, [75:07] being a venture capitalist, of all things, [75:08] what's the point about being negatively upset [75:10] about a new technology? [75:11] And in particular, it feels like it requires [75:13] high openness where, there's something about early crypto [75:17] where it attracted folks like Balaji [75:21] where there was all these grand pronouncements [75:23] of like, "Oh, Bitcoin will supersede the nation state." [75:26] It led to a lot of that kind of slightly cultish, [75:29] very cyber punk. [75:32] It really reminds me of the, [75:33] what's the John Perry Barlow letter? [75:36] "Declaration of Independence of Cyberspace." [75:38] Exactly. Exactly. [75:39] John Perry Barlow's the "Declaration of Independence [75:40] of Cyberspace." [75:41] There's a lot of that [75:42] kind of vibe about crypto— That's right. [75:44] And so it required one to be open-minded enough to think, [75:48] "there could be something here," [75:49] and I think most people are not that high openness. [75:51] High openness, and then it also got— [75:52] But you're not that high openness. [75:53] Building on that. [75:54] Well, I'm, yeah, I don't know, but I'm not introspective [75:58] so I don't have to think about that. [76:01] It also got white coded. [76:02] I think it got right wing coded, [76:03] because it got libertarian coded, [76:05] early, especially in the 2010s [76:07] when everything got politicized, [76:08] anything coded right libertarian was bad. [76:10] And then, quite honestly, and maybe [76:13] this will piss people off if I say it, [76:14] but quite honestly, if you actually want to understand [76:16] how it works, it actually is quite difficult. [76:18] It is a complex, technical thing, [76:20] and I think maybe— [76:21] People don't understand the tech. [76:22] Maybe people literally don't understand. [76:23] I dealt with this a lot when I would deal with people [76:26] who were causing us trouble in public, [76:27] and I literally would try to explain it to them [76:29] and I just fundamentally couldn't. [76:32] By the time we're using the phrase [76:33] Byzantine Generals Problem— [76:34] Yeah. Yeah, yeah. [76:35] Yeah. You're done. [76:37] It's never going to work. [76:38] My observation is that we have a friend who [76:41] talks about how crypto contains multitudes, [76:43] and that's the important thing you have to internalize, [76:45] because the criticism you will hear [76:48] is sometimes something like, "Oh, there's a lot [76:49] of scams in crypto." [76:51] That's right. And it's like, [76:51] okay, crypto is this big box [76:53] and within this big box, there's a lot of scams happening. [76:56] There's like Vitalik types [77:00] who are really interested in developing new protocols. [77:02] There's people using it as a store of wealth, [77:05] especially in emerging-market countries. [77:07] There's people who are just interested [77:09] in speculative number-go-up games. [77:10] There's people who are passionate about developing new [77:13] payment systems and they're working on [77:14] Bitcoin Lightning or something like that. [77:16] It's this big box that contains so much different stuff [77:19] and there are strengths and there are weaknesses [77:22] or there are things that we might not like. [77:24] Again, I don't like some of the rug-pulling scam aspects, [77:28] but it's just a big box with a whole [77:30] lot of different stuff in it, [77:31] and people seemed incapable of [77:34] reasoning that way. Yeah. [77:34] They'd see what they want to see, [77:36] and, plus it was also you want to see the tech guys [77:38] taken down a notch, and this is some way [77:39] that tech guys are manufacturing magic money, [77:41] and then maybe another more focused way [77:44] of what you're saying is every new form [77:47] of financial technology associated has been [77:49] historically associated with some form of bubble and crash, [77:52] and scams along with that, [77:53] and the classic example of that [77:54] that I think is illustrative for crypto, [77:57] the invention of paper money. [77:58] Mm-hmm. John Law. [77:59] John Law invented paper money in France [78:01] about 360 years ago, and it immediately sparked [78:04] what became the South Sea bubble, [78:05] and actually, he ended up, [78:07] basically, his life did not go well [78:08] after that, because people got— [78:09] No, he fled [78:10] to Venice, I think. Fled to Venice [78:11] and basically died poor, and— [78:13] High-yield finance may be another example. [78:14] What's that? High-yield finance. [78:16] Michael Milken. Oh yeah, yeah, yeah. [78:17] Junk bonds. Junk bonds. [78:18] Junk bonds were completely discredited [78:19] by the time Mike Milken was sent to jail. [78:21] Yeah, junk bonds had been completely discredited. [78:23] because everybody, again, the moral story, [78:24] everybody knew that it led to this massive bubble [78:26] of all these deliberately high-risk bonds. [78:32] Who would ever do that? [78:34] That decade later that, that market was much larger [78:36] than it ever had been in the 80s [78:37] and was extremely well respected [78:39] and it played a huge role in the [78:40] build out of everything since, [78:41] and so, new kinds of money lead to— [78:43] So— New kinds of scams, [78:44] lead to new rebels— Speaking of new [78:46] kinds of money. Need new. [78:46] Yeah? [78:47] How do you think about stablecoins? [78:48] Yeah, yeah, so stablecoins, I would say, [78:51] primarily what I think about it is [78:52] it's been super helpful to have stablecoins succeed, [78:55] because it's just an obvious incredible use case. [78:57] It's worked incredibly well. [78:59] They're being used all over the world [79:01] for many different reasons. [79:02] The numbers are now extremely large. [79:05] I think it's great. [79:05] It was, originally, you guys probably know, [79:07] it was originally part of Vitalik's early work. [79:10] He had a very unfortunate, [79:11] do you remember the original name for stablecoins? [79:13] No. Colored coins. [79:14] Oh, yes! Yes. [79:16] Only an ESL speaker would pick that name, [79:19] but the idea was a crypto token wrapped [79:21] in a real world asset, so that was part [79:23] of the original thinking on all this stuff. [79:25] It's worked incredibly well for dollars. [79:27] I believe that'll work incredibly well [79:29] for many other kinds of assets. [79:31] It's great, now having said that, [79:32] the crypto purist natives are like, [79:34] "Well, that's not the main thing," [79:36] because it's a bridge technology of the old world. [79:40] I think it's great and I think it's fantastic [79:42] that you have such a successful use case. [79:43] So FinTech has generally not produced great companies [79:46] or giant companies, because it's been [79:48] country-by-country demarcated— [79:49] He's here in our pub. [79:50] In fact— [79:51] He's really upset about the waiter. [79:55] You end up with these very mediocre companies [79:57] like Stripe, but fantastically managed. [80:00] It seems like stablecoins could lead [80:02] to the global scalability in FinTech [80:06] that has been the prerequisite [80:07] to making super-valuable tech companies. [80:10] We've had some FinTech ones that we're very proud of, [80:12] including Stripe. [80:13] Just, it's the level of, [80:16] one is regulation, and then it's particularly— [80:18] And the payments is different because they did go global— [80:20] They did go global. But not many companies [80:22] have been able— But they have regulatory [80:24] constraints there as well, [80:25] and then the last decade in particular, [80:27] a lot of the western countries, [80:29] they've been, if anything, on a crusade against any kind [80:31] of financial innovation, just on general principle, [80:33] and so there's been these real regulatory government [80:35] headwinds, and then just look, [80:37] dealing with the banks, [80:39] dealing with the credit card companies, [80:41] dealing with these... [80:42] They're not psyched at the idea of some kid [80:45] with some new idea. [80:46] They're just not, and so even if you have [80:48] regulatory clearance, can you actually implement the thing [80:51] is an open question, [80:52] and so, I just think there's a lot of glue, [80:55] a lot of stickiness. [80:56] And then, look, to be fair, [80:58] you could also say, look, it's a high hurdle [81:00] to go to a consumer and to say you should trust your money [81:03] with some new company, [81:04] so there's a whole issue there, so, yeah. [81:07] Yeah, so, yeah, I agree with your optimistic point of view. [81:10] Yeah, this was always part of the crypto philosophy, [81:14] which was programmable money. [81:15] If you have programmable money, [81:17] then all of a sudden you could have financial [81:18] services work a lot more like software. [81:20] You could have a much higher rate of innovation, [81:21] and you're right, maybe we're starting to get there. [81:23] Was there someone who really got you into crypto? [81:25] I would say the main person was my partner, Chris Dixon. [81:28] It was very early and had figured it out, [81:29] and then we were involved in Coinbase early on, [81:31] and so Brian and Fred at Coinbase [81:33] were super helpful in helping us understand it. [81:35] And how do you think— [81:36] Oh, and, sorry, and the Balaji. [81:38] Actually, Balaji's at the head of that list. [81:40] And how do you think Chris cracked it so early? [81:42] So, Chris' entire life has been this pursuit of, [81:47] it's just how he thinks. [81:48] He's just born to do this, [81:49] and it's in pursuit. [81:51] He uses these terms. [81:52] He used one term, he says, "What nerds do [81:53] on nights and weekends," is one way to look at it. [81:57] Second way to look at it is "Good ideas [82:00] look like bad ideas," and then his third [82:03] most recent version of that is "internet cults." [82:06] It's if it has like a thriving subreddit, [82:08] then like something's going on. [82:09] It's the other side of the people's negative emotion [82:12] on this is, the things that become movements early. [82:15] The internet enables movements. [82:18] Yeah, this is the home-brew computer club— [82:20] That's right. Thing. [82:21] That's right. John, how did you think [82:23] about stablecoins for you? [82:25] It's funny, when you were saying crypto [82:26] is an internet cult, we find that it's very vibes-based [82:29] in a funny way where there's always a thing of Stripe is [82:34] pro-crypto, we're super excited, Stripe is anti-crypto, [82:37] not going to make it, Stripe is pro-crypto again. [82:41] That's never how we've conceived of it. [82:42] We just want to build things that people find useful, [82:44] and the Bitcoin Whitepaper dropped in 2008, [82:49] I want to say, and Stripe was founded in 2009, [82:51] and so we've been watching all along, stuff. [82:54] Wasn't it February '09? I think— [82:56] It might have been '09. You're right. [82:56] Yeah, but anyway, it was just before Stripe, [82:58] and so, we've just been trying various things. [83:00] We funded Stellar in the early days. [83:02] We tried Bitcoin support. [83:03] Original Bitcoin was a horrible payment method, [83:05] you know what I mean? [83:05] And the thing we have really noticed [83:07] that's really striking is there's a level of [83:10] consumer adoption and familiarity [83:13] that allows for mainstreaming. [83:15] We just worked with Shopify, [83:16] they now offer stablecoin payments [83:19] on all of their checkouts, or they're rolling that out [83:21] on all their checkouts. [83:22] That's just not a thing that would've made sense [83:24] even three or four years ago, [83:26] and so, we're talking about the internet stuff, [83:30] just at a certain point, Google and Facebook [83:32] and all these companies start to work, [83:33] and if you try to launch Facebook in 1998, it doesn't work, [83:36] because there aren't enough internet connections. [83:39] I think there weren't enough wallets [83:41] for a lot of things to work. [83:42] Even, you look at the stablecoin supply charts, [83:45] we're growing at 40%, 50% year over year. [83:47] It's the grains of rice and the chess board. [83:49] You don't need that many years of 40% to 50% [83:52] year-over-year growth before it really works, [83:54] but it's been really striking for us over the past 18 [83:56] to 24 months where we've been trying [83:58] to make different things work at various points, [84:01] and we're going to be shut off products that don't work, [84:04] but now, all of the products are really working all at once. [84:09] Okay, I had some questions [84:09] on the Andreessen Horowitz's business. [84:10] Yeah. Why aren't you a hedge fund [84:11] in that you, or why don't you do public investing? [84:14] You don't have to be a hedge fund, [84:15] you can just do long only, but aren't you in the business [84:18] of predicting tech trends and evaluating companies? [84:20] After having done this conversation, [84:21] we think you might be quite good at it. [84:22] Exactly. [84:25] We've considered it, and it's just if you guys spend time [84:28] with public-market investors, they just have a very [84:30] different motion than what we do, [84:32] and so they just— [84:33] But is that tradition or is that fundamentally [84:37] intrinsic to the ontology of the job? [84:38] I think there would be a different way [84:39] to run public money in a way that, for example, [84:41] would've caught a lot of the mega seven. [84:42] I think that possibility exists, [84:44] and literally, you could just say it's as simple [84:45] as apply the venture mindset to to mega caps, [84:48] and away you go, [84:49] and obviously, we now know venture scale returns [84:52] when you get that right. [84:54] I would just tell you, one of the things that saves venture [84:58] is that we're locked up, and our investors are locked up. [85:02] It's a feature, not a bug. [85:02] It's an incredible feature, [85:04] and in traditional finance theory, they always tell you [85:06] illiquidity is a deficit, and it turns out it's— [85:08] Which is true, but human nature is a bigger one. [85:11] Human nature is— [85:12] Liquidity would be a feature if we were— [85:14] Correct. Less messed up. [85:16] It is so incredibly hard to no get sucked into [85:18] the psychology of the moment, [85:19] and I spend a lot of time at our firm [85:21] trying to get people to not be sucked up [85:22] in the psychology at the moment. [85:24] So, for example, it's just like an absolute ban [85:25] on television news in the office. [85:27] No, if it's on CNBC today, it does not matter to us. [85:30] If it does matter to us, we made some horrible mistake [85:32] eight years ago that we can't fix now anyway, [85:34] and if it's anything else, [85:34] we shouldn't be paying attention to it. [85:36] because the whole point of this is things that are [85:38] going to take five or 10 years in the future to develop, [85:40] and people just need to get back to work. [85:44] Okay, here's a very pragmatic challenge. [85:46] You're running public money with a venture strategy. [85:48] All right, what's your lockup? [85:49] Okay, now you got a quarterly lockup, [85:51] congratulations, big guy. [85:52] The market rips your face off, all your investors redeem. [85:55] So much for your strategy. [85:57] And so, that's just really hard, [85:59] and then you people who have gone out [86:00] to try to raise money on longer lockups [86:01] are like, "Well, why would I do that? [86:02] Illiquidity is a problem. [86:03] Why would I lock up an Apple position? That's insane." [86:06] And again, you can say, the fact that nobody did that [86:09] is illustrative of how difficult it is now. [86:11] I don't know. Maybe at some point we should. [86:13] And then the other is just flat-out opportunity cost. [86:15] which is, are you really going to spend the time [86:17] dealing with that you could be spending [86:19] meeting the next Mark Zuckerberg? [86:20] You invest in companies that succeed and then go public. [86:25] Can I tell the actual story? [86:26] So we almost did this. Yeah. [86:27] We almost started the thing [86:28] and we're like, "All right, we have the venture mentality, [86:30] we have the thing, but because of how [86:31] the public markets work, we need a public market." [86:33] I mean, somebody with some public markets background [86:35] to even be able to raise the money. [86:36] So we ran a long recruiting process [86:38] and we got down to the final candidate [86:40] and we met with him during COVID in, [86:43] I'm going to say September '21, something around that time, [86:46] and we said, "Look, just do the workup [86:51] and bring your best idea. The one company that you would [86:54] commit the portfolio to." [86:55] Would you like to take a guess what it was? [86:56] No. Peloton. [86:58] Oh, my God. [87:01] Which then proceeded to fall 99.9%. [87:03] So you, yeah, missed bullet. [87:05] Right, and by the way, at the time, Peloton, [87:07] and you remember, [87:08] we used to talk about this at the time, [87:09] we were, "Peloton, oh, this isn't just a bike company, [87:13] this is a movement." Yeah. Right. [87:14] "This is a cult, and this is a brand, [87:15] and this is immediate," and everybody had their [87:17] theory on subscriptions and recurring revenue, [87:19] and, you know— Well, during COVID, [87:19] where people overestimated the permanence [87:21] of the behavior changes. [87:22] Yes, exactly, well, there was that, [87:24] but there was also just that kind of company, [87:28] fitness is a trend, fad-driven business historically, [87:32] and so anyway, that just felt like a message from God. [87:38] Go back to public market investing. [87:39] So, you invest in companies that then go off and succeed [87:41] and go public, like Coinbase or Airbnb [87:44] or all these sorts of companies. [87:45] You then, because they're public, [87:48] you get to distribute the stock [87:49] and so you distribute it to all the LPs. [87:51] They get their shares. You get your shares. [87:54] Yeah. Do you hold the companies? [87:56] Do you make a decision? Is it formulaic? [87:58] Is it not formulaic? [87:59] Are you secretly a public-markets investor [88:01] because you have to make these decisions? [88:02] Yeah, so, to be clear, there's two parts to that. [88:05] The part, each of us as individuals do whatever we do [88:07] with the stock. Yes, but what do you do? [88:08] What do I do? [88:10] Nothing specifically, but I'm basically saying, [88:11] do you make active decisions [88:12] or is it totally formulaic? [88:14] Well, lemme tell you how we do it as a firm, [88:15] and then it gets to the individuals. [88:17] As a firm, we try to make it as mechanical as possible. [88:20] We're trying to get out of the psychology [88:22] of whatever's happening at that moment. [88:23] So you try to define a process upfront, [88:25] but you do want to be discriminating, [88:28] and so we have a magic box formula [88:30] of things like, are the founders still running the company? [88:34] Quality of the founders, are they beating their numbers? [88:39] What's the growth rate? [88:40] What's the second derivative? What's the service [88:41] like in the pub? [88:41] Exactly. [88:43] Do they tolerate low-performing bartenders? [88:46] Yeah, and then, yeah, we have some schedule against that. [88:49] There is a theory afoot [88:51] and Sequoia is pursuing it, that basically [88:53] the venture firms and their LPs have left [88:56] enormous amounts of money on the table [88:57] by distributing too soon, [88:59] and, if you backtest over 50 years, [89:02] the best strategy, at least for the top firms, [89:04] probably would've been to hold everything in perpetuity, [89:06] and so, Sequoia, notably, is trying a strategy [89:08] where they're trying to do more of that. [89:10] I will tell you the LPs don't like that. [89:13] The LPs— [89:15] The LPs just want their shares. [89:16] Of money in and out, and, yeah, [89:17] and they do have a positive argument [89:19] that says, "Look, we're not paying you [89:20] to manage public money." [89:21] Yeah. And, by the way, [89:22] they have their own needs, [89:23] and, by the way, they have their own needs [89:24] now more than ever. [89:26] They're under real pressure in a lot of cases, [89:28] and so, if you ask an LP, [89:32] they will tell you, "Yeah, we want you to try [89:34] to shoot the lights out on as long-dated [89:35] a horizon as possible." [89:37] Having said that, like— [89:38] As soon as humanly possible. We want out money back. [89:40] Get us some money, please. [89:42] And so, where this comes up is [89:45] it's just the thing of, well, should we hold it [89:47] for another three years and go for another doubling [89:49] or should we bird in hand on that? [89:51] Anyway, so we try to run that mechanically. [89:53] On the individual side, it really varies by the individual, [89:55] just based on idiosyncratic life circumstances. [89:57] Off to big-company world [89:58] for a couple of questions. Yeah. Sure. [90:00] How much should big companies focus on their competitors? [90:03] So, this is a real double-edged sword. [90:06] So, the easiest thing in the world is [90:09] to focus on your competitors, [90:10] because you've got somebody to benchmark index against, [90:12] and it's just been amazing how many other big companies [90:15] start or stop their VR and AR programs [90:17] based on what everybody's doing at that moment. [90:18] They seem to have outsource their thinking entirely to Meta, [90:20] and so there is this dysfunctional version [90:23] where you're outsourcing your thought to the competitor, [90:25] and then there's the peter critique of [90:26] you're getting into these Girardian spirals, [90:30] and I think there's something to that. [90:31] Having said that, I see the other side of that all the time, [90:34] which is the Andy Grove side, [90:35] which is "Only the Paranoid Survive," [90:37] and isn't it great if you have an intellectual framework [90:41] to be able to not think about your competition? [90:43] because that's a lot more fun. [90:44] Yeah. If your competition's good, [90:46] thinking about them is actually really painful. [90:48] If you have this enlightened point of view [90:51] that says you don't ever have to think about them, [90:52] you're letting yourself off the hook, [90:54] and so I think there's- [90:55] Maybe the answer is whatever's most painful thinking about [90:57] them or not thinking about them, is best. [91:00] Well, and this gets to what I've experienced [91:01] with big companies. [91:03] What I've experienced with big companies, and, by the way, [91:06] this includes, in a lot of cases, fast-growing startups, [91:08] they think a lot about their competitors [91:10] for the purpose of trying to basically, [91:12] essentially ultimately copycat with their competitors. [91:15] If your competitor's decent, for whatever it is they do, [91:18] you assume they must have some [91:19] analytical reason they're doing it, [91:20] and so there's this natural tendency to try [91:23] to build the analytical case to do the same thing, [91:24] and so, there's an over focus in that way. [91:27] Having said that, I can count the number [91:29] of true competitive teardowns. [91:32] Yep. I don't know, [91:33] maybe on one hand, that I've ever really seen, [91:36] because again, your pain point, [91:37] the most painful thing in the world is to talk honestly [91:39] about somebody beating you. [91:40] Yeah, I always find that Jeff Bezos, [91:42] we're not competitor-focused, we're customer-focused, [91:45] a bit of a misdirection, because, again, [91:46] at least at Stripe, we think that our customers [91:49] are very smart, and so if they're picking something else, [91:52] that is some signal of revealed preference [91:55] that a well-informed person trying to do [91:57] the best thing for them says, "This is better than Stripe," [92:00] and so, we do a lot of secret shopping, [92:02] we do a lot of tearing down. [92:03] We want to understand what's out there, [92:04] and again, as you say, that shouldn't define the roadmap. [92:07] You should be able to come up with your own products, [92:08] but if you're not coming at it from an informed place, [92:10] something is horribly wrong. [92:11] I think, yeah, it's some combination of [92:13] you need to be brutally honest with respect [92:15] to what your actual issues are, [92:16] and those actual issues include [92:17] you're losing for reason XYZ. [92:20] I mean, in some ways, what they're saying [92:21] is bureaucracies avoid pain— [92:22] Yes. And so you need to [92:23] steer them into pain. [92:24] But I would say it slightly differently, [92:25] which is I have found people willing to tolerate any level [92:27] of chronic pain in order— [92:29] to avoid acute pain. Avoid acute pain. [92:30] In order to avoid acute pain, [92:31] and so people would much rather lose slowly [92:34] over five years, than have the conversation [92:36] that involves a dramatic change to stop losing. [92:39] Wow. And I've seen that [92:40] over and over again. [92:41] It's almost impossible to get people to do that. [92:43] Just, the level of aversion is incredibly high. [92:46] What founders or companies do you respect— [92:48] People seem fine just bleeding out. [92:50] It's just incredible. [92:52] You see it in politics. [92:57] I don't want to name names, [92:58] but there are political parties, let's say, [92:59] in various places around the world, [93:00] where you just look at it [93:01] and you're just like, "I can't believe that you're [93:03] willing to inflict this strategy on yourself [93:05] with these results that are clearly not working." [93:07] Actually, that leads [93:08] to a next question. And yet, they will not [93:09] revisit their core assumptions. [93:10] If you look at companies that have died [93:12] over the last 20 years, they do seem to have [93:14] these very long operatic deaths, [93:18] and they change less than you would think. [93:20] Yes. Do you think that's because [93:21] of people that are prescient and see it, just exit? [93:24] Yeah. And so, you've got [93:26] a selection effect in the people that remain, [93:28] or is it just that it's too socially awkward [93:30] to have a conversation that says, [93:31] "We've backed out"? Most people would rather [93:34] just put one foot in front of the other. [93:36] Most people don't want to rock the boat. [93:37] Most people don't want to be the skunk at the garden party. [93:40] Most people don't want to call their own baby ugly. [93:44] Yeah, most people don't want the reputation [93:46] of being a troublemaker. [93:49] It's this very interesting signal [93:52] you have to decide whether you want to send as a leader, [93:53] which is do you want people to bring you bad news? [93:55] Because, if all people are doing to you every day [93:57] is bringing you bad news, number one, [93:58] you're going to slit your own wrists, [94:00] because that fucking sucks, and then, number two, [94:01] you don't want people to just be complainers. [94:03] Yeah. Right? [94:04] And so, maybe the more advanced version is [94:06] only bring me a problem if you're [94:07] also bringing me the solution. [94:08] But, okay, now your life as a CEO is better, but what if [94:11] there really is a problem? Is a problem, [94:13] and they don't— And they don't have [94:14] the solution. Because it's beyond them. [94:15] Yeah. And it's beyond them, [94:16] and then they're the one that you're going to give [94:17] the negative performance review to, so... [94:19] By the way, the other twist on [94:21] the big company failing thing, [94:22] which I think is really underrated, [94:23] is the big companies that fail, [94:25] the way the story gets written is they never figured it out, [94:28] and the easy example of this is always Kodak, [94:31] for example, they never figure digital photography. [94:32] What you often find in the backstory is, [94:33] no, they actually figured it out and they did it too soon. [94:36] That's interesting. Kodak had actually [94:37] a very active digital camera program [94:39] before— Then they got burned. [94:40] And then they got burned. Once burnt, twice shy. [94:41] Twice shy. [94:43] By the way, Yahoo had mobile early. [94:44] Yahoo was all over mobile between 2002 and 2006, [94:48] and then they got burned so hard on it [94:50] that by the time the iPhone appeared, it was too late. [94:53] Yeah, I think that if you did WAP, [94:55] you were unlikely to succeed in the post-iPhone world. [94:57] Yeah. Yeah. [94:58] And, quite frankly, well, you mentioned [95:00] the big tech companies, a lot of the big tech companies, [95:02] they had internet fully deployed internally, [95:05] they had TCP/IP products. [95:06] They actually knew it quite well. [95:08] They were running it, it just was something [95:10] that they were very used to that they didn't [95:11] really think about in a new way, [95:12] and, yeah, there's the status quo bias thing. [95:15] So, this is a good segue— You had very [95:17] intelligent-sounding reasons as to why it won't work [95:19] from a recent document. Oh, people are really good [95:23] at the analytical explanation, [95:25] either as to why something won't work, [95:26] or conversely, why something is going to work [95:28] when it's currently failing. [95:29] Yeah. No, but again, you just [95:30] get it to sound very convincing, where it's like, [95:32] "That's a great point. [95:33] We actually tried that 18 months ago." [95:34] That's right. Yes. [95:35] It is, no man steps in the same river twice, and so... [95:40] That's a good segue into you've been on many boards. [95:42] What makes a good one? Or maybe what makes a bad one? [95:46] Yeah, I mean, step one is if it's a successful company. [95:49] Step two is— [95:51] Which way does the cause and effect? [95:52] Step two is if it's a good CEO. [95:55] Just practically speaking, the boards just [95:56] can't do that much, and even the old cliche [95:59] is to hire and fire the CEO, [96:00] and even that, it's really fraught with peril. [96:02] Yeah. It's very easy for a board [96:03] to blow that up. [96:04] By the way, again, it's often— [96:06] I do remember your blog had a, [96:08] "How do I hire a professional CEO?" [96:09] And the answer was one sentence, [96:10] expecting this long article, and it's like, "Don't." [96:12] Don't. If you need to do that, [96:13] sell your company. Sell your company. [96:15] And that's probably an overstatement. [96:18] There have been some very successful professional CEOs [96:20] over the years, John Chambers and Frank Slootman and others, [96:22] but, yeah, look, it's just really hard. [96:25] Is the company going to succeed or not? [96:26] Is the CEO great or not? [96:28] Is the company on the right side of history or not? [96:30] That's honestly most of it. [96:31] The board— Well, do you think [96:32] boards matter then? [96:32] It's one of those things, you can't not have one, [96:35] which is, if you're on another board, [96:38] then you're, as a CEO, legally liable [96:40] for every screwed-up thing that happens. [96:43] You're much more likely to go to jail. [96:46] You're much more likely for things to spin out of control. [96:47] There are real requirements. [96:49] Governance needs to be taken seriously. [96:50] You're representing a lot of other people's money, [96:52] so there's that, and then do you want to have [96:55] absolute dictatorships with no examination inside ever? [96:59] And then, aspirationally, obviously, [97:01] the hope would be to be able to positively contribute. [97:03] Yeah, you're giving the governance explanation [97:05] and you're saying that it's rare [97:06] that founders are actually removed, [97:08] or CEOs are actually removed, [97:10] and then even the cases where they are, [97:12] maybe things are too far gone and everything, [97:13] and sure, maybe that's true, [97:15] but I feel like I would make a cultural pitch [97:17] where, let me try this on and you can react to it, [97:20] we've found the Stripe board very useful [97:22] because it's important to have to organize your thinking [97:25] and have some accountability mechanism [97:26] where you go on a quarterly basis and talk about things, [97:28] and then, we're doing this for the first time, [97:31] and so there's lots of people on the Stripe board [97:34] who have a different set of experience and come to us [97:36] and advise us on various things, [97:38] and we've gone and tried to pick the hall of fame [97:42] of various industries who can then go opine on things. [97:45] And I actually notice, [97:47] when I talk to way-earlier stage founders, [97:49] I think they underrate the value of a good board, [97:53] where they are worried about the governance thing [97:55] you say, where they don't want to give up a whole bunch [97:57] of board seats and then have to do management [97:59] of VC personality and everything, which is true, [98:02] but they don't seem to take seriously, [98:04] and, again, maybe they just get this from investors, [98:06] but they don't seem to take seriously the idea [98:08] that you can put together a group [98:12] who will meaningfully increase the odds of success [98:15] of the company. [98:17] I don't know, is that just a particular thing to us? [98:19] We needed more help than others [98:21] or would you agree with that broadly [98:23] as a cultural explanation, where they're pretty useful [98:25] culturally for management? [98:26] Yeah, so what you just said is what we aspire to, [98:29] and what we aspire to is that the boards [98:30] that we're on are like that, and that the CEOs [98:32] that we work with want to have a board like that, [98:33] and that we're able to be a contributor to it, [98:35] and so we aspire to that. [98:36] I think there are many examples of that being true, [98:38] and hopefully, on that, I've been an example of that myself. [98:42] I think that's all true. [98:43] Having said that, I guess, a board cannot rescue [98:45] a failing company. [98:46] Sure. Well, yeah, [98:48] but there are a lot of people [98:50] on a lot of boards in a lot of companies [98:51] that are failing, that are spending an [98:52] enormous amount of time trying to rescue those companies, [98:56] both in and outside of tech, [98:57] and so, it's just a high-order bit is still succeeding [99:00] or failing, and it's still quality people versus not. [99:02] It ties into your, [99:04] someone once said— The easiest thing [99:05] in the world is to go on the board of a company [99:06] that is going to succeed wildly no matter what you do, [99:09] and then to take credit for it after the fact. [99:11] I mean, that sounds fun, but— [99:13] Having been through it, [99:15] the hardest thing in the world is to be on a team, [99:16] on a board, where you're struggling valiantly [99:18] to keep the ship from going down. [99:20] And the ship is going down. [99:21] So, that goes back to [99:22] can you hire— I've been on those too. [99:23] Can you hire great CEOs or are those great CEOs, [99:26] someone once described it to me, [99:28] the people that have reputation for great professional CEOs [99:30] are actually great stock pickers. [99:32] Yeah. They understand tech [99:33] deeply enough that they pick the company— [99:35] Right. That's in a great position. [99:37] Same thing for VCs. So— [99:37] Same thing, yeah. Right. [99:38] You can't hire them to turn around a failing company, [99:41] because they self-select out of it. [99:44] I don't know, there's exceptions to everything. [99:46] Every once in a while you get— [99:47] Actually, that is also a great point, [99:48] which is there's a world full of heuristics in VC, [99:51] single founders, multiple founders. [99:53] Yeah. But, there's so many [99:54] exceptions to each rule— You'd never back [99:57] a married couple, and then you didn't back Cisco, right? [99:59] Alright, last call, but you just made it. [100:01] That was the saddest [100:03] pub service. That was the [100:04] slowest pint. [100:06] Is that the Irish— [100:07] Here, let me get you a cup of coffee. [100:08] Is that the Irish standard? [100:11] Does this explain, or are you actually a...? Yes. [100:14] We've all seen the productivity stats. [100:16] We specialize in economic infrastructure, [100:17] not in hospitality service. [100:22] I would just like to note that the result [100:24] of that entire exercise is one less of [100:26] my cups on the table. [100:28] Yeah, you're negative [100:30] two drinks from that interaction. [100:32] Net subtractive. [100:33] You think people under study the Elon [100:37] method for running companies? [100:39] A hundred percent, yes. [100:41] Maybe just briefly describe that method [100:42] and then why everyone is so incurious about it. [100:47] Yeah, and there's two reasons they're incurious about it. [100:49] There was the original reason they were incurious about it, [100:51] and now there's the new reason they're incurious about it, [100:53] which is Elon also generates emotion in people. [100:55] Yeah, so, look, you guys, you know, [100:57] so how do you run a company? [100:59] Well, there's been a hundred years of management books [101:01] starting with Alfred Sloan's book. [101:03] Alfred Sloan built General Motors, [101:04] and so, Alfred Sloan famously wrote a book [101:07] that people like Andy Grove learned from, [101:08] that basically said, "Here's how you build a large, [101:10] multi-national, multi-product-line industrial company." [101:11] And so there's this system, and it involves [101:13] somebody at the top of the company [101:14] that's overseeing this machine, [101:16] and fundamentally, they're getting reports [101:18] and then respond to the reports, [101:20] and then there's all these rules. [101:22] I would say, both rules inflicted [101:25] from the outside and rules generated internally, [101:28] and then there's Elon, who just doesn't do any of that. [101:30] He just doesn't do any of that, [101:33] and has a completely different playbook, [101:34] and the Elon playbook, in a nutshell, as far as I can tell, [101:37] I haven't worked for him directly, but from observing him [101:39] and working with him, as far as I can tell, [101:40] it's basically, number one, it's only engineers. [101:44] People who matter in your company are the engineers, [101:45] the people who understand the technical content [101:47] of what you're doing for technology companies, [101:49] and then you only ever talk to the engineers. [101:51] You never ever talk to mid-level management. [101:53] If you have it fine, if they need it for whatever, [101:55] to do their, whatever, vacation policy or whatever, [101:57] it's fine, but if you are the CEO, to get the truth, [101:59] you only talk to the line engineer, [102:01] and so you just ruthlessly violate the chain of command [102:03] at all times, and then your job, as the CEO, [102:07] is, every week, to fix whatever is [102:08] the most important bottleneck to the company's progress. [102:10] And the way that you do that is you parachute in [102:12] and you find the engineers that are working on that problem [102:15] and you basically stay up with them all night [102:17] until they fix the problem, [102:20] and then if there's no current major bottleneck, [102:23] you spend your time instead, doing engineering reviews, [102:26] specifically engineering reviews, not product reviews, [102:28] engineering reviews, and you get all the engineers together, [102:31] and you have them each present [102:33] what they're doing for five minutes, [102:34] and the result of that is, you know every single engineer [102:36] in the company, you know exactly what they're working on. [102:39] If somebody's not good, you fire them on the spot. [102:42] If somebody's great, you go all out to get them. [102:44] But what's the inverse of that? [102:45] because for 10 years after Steve Jobs, [102:47] we had people doing mimetic bad version, [102:50] wearing turtlenecks, trying to— [102:53] Be an asshole. Do the social style. [102:54] Exactly, I was trying to say that [102:56] more diplomatically, but yes, being an asshole. [102:58] The people were being, [102:59] not Steve. No joke, that they were big. [103:00] That was the misinterpretation. [103:02] What is the danger for entrepreneurs, [103:05] what's the bad version of copying you on? [103:07] Oh, the bad version is, this is the critique, [103:10] actually, my partner Ben levies this critique. [103:12] He's like, "Marc, the thing you don't get is as follows," [103:14] which is that, which is, that assumes you have [103:16] somebody like Elon, who can hold the entirety [103:18] of every engineering topic, [103:20] and every business topic in their head all at the same time, [103:23] and so, when you're sitting there [103:25] with the 23-year-old engineer [103:27] and you're working with them to redesign [103:28] the database architecture or whatever, [103:30] you actually are qualified to do that, [103:32] and that you're qualified to do that, [103:33] not just that one time, but every time, [103:36] and then again, this goes right back [103:38] to the last topic we just talked about, which is like, [103:40] okay, how many of those people exist [103:42] who can possibly do that? [103:43] And we know the answer is one. [103:46] Yeah. [103:47] I believe the answer is 10, or a hundred, or a thousand. [103:49] I don't know if it's a million. [103:52] I tend to think we have more of those people [103:54] than we think we do. [103:55] I see a lot of founders who struggle with this. [103:58] So my observation for how founders try to figure this out, [104:01] is, in the beginning, they run everything. [104:03] You just do everything. [104:04] You just do everything, run everything, because you have to, [104:06] and you have to have a unified vision [104:07] and you don't have this army of people anyway, [104:08] and so you just do it, and then at some point, [104:11] your high-value board comes to you [104:13] and says, "You idiot, you're micromanaging. [104:14] You need to bring in all these executives," [104:16] and then what happens is, then you go the other way. [104:18] You over delegate, and then your high-functioning [104:20] board says, "You idiot, you're not involved [104:21] in enough of the details," and then you correct. [104:24] And then, what most of the [104:25] successful founders I work with do is, [104:26] they end up with a hybrid model [104:27] where they're deep in the details on some things, [104:29] but they have a traditional system on the other hand. [104:32] And do you think that works pretty well? [104:33] I think, for most of the founders we work with [104:36] that have very successful outcomes, I think that generally [104:38] is what they do. [104:39] I think it works well, but it's not the Elon method. [104:42] Sure. It's not the Elon method. [104:43] By the way, there's other aspects [104:45] to the Elon method. I was going to say, [104:45] I feel like there's more. There's other aspects. [104:47] So another aspect of it is, the function and purpose [104:50] of the legal department is to file lawsuits. [104:51] Mm. [104:54] And I am not interested in all the rest of this stuff. [104:56] You can go deal with it, if you want to, [104:58] whatever, whatever, whatever, [104:59] but anybody who goes up against us, [105:02] we are going to terrorize. [105:04] We are going to declare war, [105:05] and then, of course, as a consequence to declaring war, [105:07] we're not always going to win all the wars, [105:08] but we're going to establish massive deterrence, [105:10] and so nobody will screw around with us. [105:12] By the way, let me give you number three, [105:13] which is becoming more and more salient, I think, [105:15] and something we're trying to get our founders [105:16] to do a lot more of. [105:17] Number three is, it's going to be a cult of personality [105:19] and it's going to be a cult of personality, [105:20] not just inside the company, but outside the company. [105:23] And we're not going to spend any money on marketing. [105:25] We're not going to put any time into IR. [105:26] What we're going to do is we're going to put on the show [105:28] of all time, and the company, and the stock, [105:31] and the books, and the videos, and the products, [105:33] and the jobs, are all a function [105:36] of the cult of personality. [105:37] I would add three things to that list too, [105:41] and you can tell me if you think you agree with these. [105:45] And, by the way, I thought the Walter Isaacson book, [105:47] it got a mixed reception, [105:49] but I thought if you want to study the Elon method a bit, [105:51] it was actually pretty useful for that, [105:54] the recent biography. [105:55] One is picking sensible metrics for the business [106:00] at any one moment in time, [106:01] and so, with SpaceX, as they were building up [106:06] the launch business, dollars per kilo to orbit [106:10] being the metric they were going to optimize for it. [106:11] That's not totally obvious that it falls out. [106:13] Even Tesla, as they're wrapping up production, [106:16] it's deliveries per week. [106:19] You could have focused on revenue, [106:20] you could have focused on profitability, [106:21] you could have focused on deliveries per year. [106:23] The number of deliveries per week [106:25] rolling off the factory line is, itself, [106:28] an interesting choice of high-level metric. [106:29] So a big focus on, I think there's lot of this in Twitter [106:32] as well, when he took it over, [106:33] focus on what are the right metrics that we should be... [106:36] And some of the criticism that's been levied at X [106:38] is their focus on engagement minutes on the site [106:41] has led to things like the ban on URLs, [106:46] the de-boosting of URLs, [106:47] which a lot of people think is pretty silly. [106:48] Okay, so, one is choosing the right metrics. [106:50] The second is creating a sense of urgency, [106:53] and people talk about this as inventing crises, [106:56] but I would say, the generous version, [106:57] is shortening the time horizons, [106:59] and so, it's funny, Elon was going around talking about [107:01] when he was sleeping on the floor of the factory [107:04] in Nevada for Tesla, that "Tesla will go bankrupt [107:08] if we don't do this, and if we don't figure out [107:10] Model 3 production." [107:12] Tesla was a $200 billion company, [107:13] by market cap, at that time. [107:14] So, it's Tesla will go bankrupt [107:15] or do a very, very non-dilutive equity rate, [107:19] but creating a lot of urgency around this idea [107:21] of fixing production and sleeping on the factory floor, [107:23] which clearly shortens the timeline. [107:25] And then the third is actually, [107:26] the business is already capital efficient, [107:27] so I'm curious if you see this with hardware companies. [107:29] I think sometimes hardware companies [107:31] can be really indulgent with capital [107:34] where they say, "Venture capitalists will fund my vision [107:36] of exploration for five or 10 years," [107:39] and the risk now, is people get into robotics [107:41] and stuff like this, and then you get the self-indulgence, [107:42] and it's like, "I will do my science project for ages [107:44] and then I'll maybe figure out a product and figure out how [107:46] to commercial it." I would say the other thing [107:47] hardware founders do is, they fall in love [107:49] with the hardware and the product. [107:50] Yes. And they can almost [107:52] redefine themselves as like producers of science, [107:55] or beauty, or product, and forget they're [107:57] running a business, or even worse, [107:59] start to think of running the business [108:00] as slightly unpleasant— An annoyance. [108:03] Beneath them. Exactly. Yeah. [108:04] And maybe even not intellectual enough. [108:06] Right, and so, Elon's companies have always been [108:09] very capital efficient and build a bad one [108:11] and then build a good one, [108:11] and so The Boring Company bought a commercial tunnel-boring [108:14] machine before they started developing their own. [108:16] Tesla had the master plan [108:17] where they build the low-volume roadster [108:18] before they get to the high-volume stuff. [108:20] SpaceX just for what they do, has never actually burnt [108:22] that much capital lifetime, and got grant money, [108:25] they were selling to the DOD, all this kind of stuff, [108:27] and so, yeah, would you agree with those three? [108:30] And do you think people can pick and choose? [108:32] because we can take some of those things without maybe [108:34] the lawsuit department or something. [108:37] Yeah, so, I think that's all right. [108:38] I would maybe add one more thing [108:40] or distill it out of a bunch of these, [108:41] which is basically truth seeking at all cost. [108:44] Mm. That's a good point. At least, I find this [108:46] to be the case with him, and I think this is really not, [108:49] especially people who are mad at him, [108:50] really don't understand this. [108:51] He really, really genuinely wants to know ground truth, [108:53] and he really genuinely does not want [108:54] to know anything that's not ground truth. [108:57] And again, it goes back to our thing [108:58] of how to confront bad news. [109:01] He's absolutely ruthless and relentless [109:03] in making sure that he actually understands what's going on, [109:05] and you would think that that's common, [109:07] and I've not found that to be common at all [109:09] among people in business. [109:11] Or, related to another thing, which is, [109:13] you mentioned the thing where, literally, with Elon, it's, [109:16] "We're all going to die, if we don't get this rolling." [109:19] Every other typical startup founder, [109:21] me when I was doing it, [109:23] you get in front of the company, you're trying to— [109:24] Put on a brave face. Optimistic, brave face. [109:26] "It's going to be great. You should have faith. [109:28] You shouldn't quit and go to another company," [109:30] like, "Please, stay with us. You're going to be great," and— [109:32] Trying to weed out the non-believers or something? [109:34] Apparently, yeah, and I think it's urgency, [109:38] but, yeah, literally, it is just to be the guy [109:41] who can show up there and just be like, [109:42] "Yeah, if this doesn't happen, it's going bankrupt." [109:45] The number of other companies where that would happen, [109:47] okay, the talent would just bleed out, [109:49] and then, maybe I could add one more thing to this, [109:52] which is he has what's, you mentioned Steve, [109:54] he has what Steve had, which is the people who work for Elon [109:56] and the people who work for Steve, [109:57] they often report after the fact that they did [109:59] the best work of their lives, [110:00] and they often report that they could have [110:03] had difficult interactions along the way, [110:06] or they could have had whatever, whatever, [110:07] or, by the way, maybe it didn't even end well. [110:09] Yeah. And— [110:11] And they were pushed. [110:12] Yeah, but literally, they'll say like, [110:14] "Wow, I got to work on the iPhone." [110:16] There's a lot of very good ex SpaceX founders. [110:18] Yeah. Right. And they imbibe [110:21] a work ethic that reminds me of, I don't know, [110:23] Goldman Sachs in the 1990s or something, [110:25] where they work incredibly hard, [110:27] and they think from first principles [110:30] and they're truth seeking. [110:31] Yeah, that's right. And they're risk taking, [110:33] both technically and, [110:34] they're risk seeking technically, [110:37] and risk avoiding in business. [110:38] Yeah, so then my version of your question is, [110:40] I call this the question of the milliElons, [110:43] it's like, okay, if a full Elon is a thousand milliElons, [110:47] right? [110:49] You can microdose. [110:50] Yeah. Can you microdose? [110:51] So, can you operate at the level [110:52] of a hundred milliElons, or at 10, or at one? [110:56] And a huge number of observers of Elon, [111:00] it's a classic thing. [111:01] He gets the classic feedback. [111:02] Steve used to get this feedback. [111:03] Lots of people get this feedback. [111:04] Just, "Wow, you're great. If you could only just do 80%, [111:08] if we could just get the 800 milliElon version [111:10] and you could just not do the other 200 milliElons, [111:15] you'd be so much better," [111:16] and literally, what I found with these guys is, [111:19] they've heard that a thousand times [111:20] and it's a completely no-up of a statement, [111:22] because, for them, there's no reduced version. [111:26] Yeah. And so, if there's [111:27] no reduced version of it for them, [111:29] is a normal person going to be able [111:31] to construct like an optimally-titrated [111:34] dosage of milliElons? [111:36] And, I aspirationally believe that you should be able [111:39] to learn things and replicate, [111:41] but it is a system. [111:45] It's not just a set of practices. It's an entire worldview. [111:49] I'm not sure it's a whole system [111:51] where if you don't have one thing, [111:53] the whole thing falls apart. [111:55] I feel like you can— [111:56] Well, the other part of that, though, [111:58] that would be one, the other way of looking at that, though, [112:00] is the person capable of doing the partial version? [112:03] No. That I— [112:05] You see what I'm saying? Yeah, that I can buy. [112:06] Are there people— [112:07] Yes. Who can do [112:08] the 300 milliElon version of it? [112:10] Yeah. Yes. Maybe. [112:13] I wish I had met more of them by now, [112:16] and then the other side of that is why is it understudied? [112:18] And, literally, I think this goes back to the same thing as [112:20] why do people get mad about cryptocurrency? [112:22] Is, I think, this— [112:23] It's tribalism. Yeah. [112:25] There was always something about him [112:27] and how he operated that caused people [112:29] to have an emotional response, [112:30] and then that is now magnified a thousand x [112:33] or a million x, and people are just not having it. [112:36] And part of it he's polarized the market very deliberately, [112:44] in the same way that I think a lot [112:44] of great entrepreneurs do, which is unique. [112:46] People tend to either love him or hate him. [112:47] They either love the products, hate the products. [112:49] That's very helpful from a business standpoint, [112:51] recruiting standpoint, because it does create [112:52] this cult-like thing. [112:54] The thing you don't want in any market [112:56] is a lack of differentiation. [112:57] He a hundred percent always has that, [112:59] but as a consequence, I believe there are a lot of people [113:02] who should be learning a lot more from him [113:03] who cannot bring themselves to do it, [113:05] and to their own detriment. [113:07] Can I talk about the media? [113:08] So, I feel like my framework is that there are [113:11] often these new technologies [113:13] that then cause an explosion in interesting media activity [113:17] in new companies and things like that, [113:18] and so there was the cable boom, [113:22] and I'm excited for John Malone's, [113:23] new book, but I saw an interview with him recently [113:25] and he was talking about, they just thought up [113:27] a lot of new channels, when they had this pipe [113:30] going into people's homes that could support [113:32] a lot of programming, they had to invent [113:33] new programming for it, he was talking about [113:35] creating Fox News, because they were like, [113:36] "Well, the existing channels seem a little bit to the left, [113:39] and conservative talk radio is really popular, [113:42] so it seems like conservative news channels [113:44] should work really well," and it did. [113:46] So that was cable, then the internet came along [113:48] and famously, really worked from a media perspective, [113:51] and in particular, that was the big nail in the coffin [113:53] for local newspapers, where they were the [113:56] main distribution outlet to people previously [113:59] for information, and the internet went over the top. [114:02] I feel like plausibly, X is a big enough change [114:05] to be a new media platform. [114:07] A slightly trivial example, but TPBN [114:10] is kind of a CNBC competitor where, I saw Mati [114:13] from ElevenLabs, a great a16z company, [114:16] and they did a fundraise [114:18] and he went on TBPN to talk about it, [114:19] but previously, that would've been CNBC, [114:22] but now TBPN is where he chose to go. [114:25] And that's one example. There's lots of others. [114:29] Is X that big a deal from a media perspective [114:31] as to be cable, the internet, then X, [114:33] or am I missing something? [114:35] I think it is, and maybe the twist I would put on [114:36] the TPBN or the cable thing is, [114:40] actually, this is also, what I'm about to say, [114:41] a big deal in sports. [114:43] There's also now the clip, [114:45] and clips used to be weird and esoteric, [114:47] and now clips are the main way that people consume content. [114:50] I see, so X and short-form generally. [114:52] Exactly, yeah. Yeah. Mm. [114:53] And so, for example, a TPBN episode, [114:56] or for that matter, a sports game, [114:58] now generates five or six or eight clips, [115:00] or an interview, or, hopefully, this discussion, [115:04] and then those clips go hyper viral, [115:06] if you're doing it right, [115:07] but it's very common when you look at the analytics [115:09] that the clips get a thousand times the distribution [115:11] of the actual program itself. [115:12] Interesting. And so, I think [115:14] there's this art form, it is one of the reasons [115:16] why a lot of historical television shows [115:17] never figured out what to do with the internet, [115:18] because they didn't really understand [115:19] that the internet native artifact was the clip. [115:22] But the new media entrepreneurs, [115:25] I think, tend to really understand that, [115:26] and so, yeah, I think that's true. [115:28] Having said that, the impact of the internet [115:29] is still mostly what it's been this whole time, [115:31] which is a disintermediation mechanism. [115:34] In the cable era, there were only 200 channels— [115:37] Right. Or whatever it was. [115:38] In the internet, there's a billion. [115:39] Yeah. So, the overwhelming trend [115:42] is still disintermediation desegregation. [115:44] Yeah, and Substack— Democratization. [115:45] Obviously, is the other big trend, to me, [115:47] in media right now. [115:48] And Substack's a great example, [115:49] because, of course, Substack as a thing [115:50] is a centralizing phenomenon. [115:53] It's a singular platform, and we have growth charts [115:56] and we're proud when they go [115:57] up that way. Unbundling and bundling. [115:58] Exactly, but it's not a re-bundling [116:01] in the form of a new magazine, and specifically, [116:03] the way that Substack thinks about it is, [116:04] they're not a publisher, there're a platform, [116:07] and the distinction is they do not have editorial judgment. [116:09] They are not trying to create bundles. They're not— [116:10] And the economics are different for the publishers. [116:13] It's land reform for journalists. [116:14] Yes, exactly. Exactly right. [116:17] But again, you would still say notwithstanding [116:18] the success of Substack as a centralized platform, [116:21] it's overall effect is still disintermediation, [116:23] because, specifically, what it's doing is it's bleeding off [116:26] many of the talented individual contributors [116:27] at Legacy Media— Yes. [116:28] To have their own Substacks. [116:30] It somehow feels to me like we're not done [116:33] with the media changes. [116:34] Yeah. I think that's true for sure, yes. [116:38] Oh, sorry, the media changes wrought [116:40] by just this latest platform change [116:42] of X and clips, the fact that, again, TPBN, [116:46] which I mention just because they're in our corner [116:48] of the tech world, is from this year? [116:50] Last year? Yeah. [116:50] It's a very new thing, and, [116:52] yeah, we haven't seen all the last changes. [116:55] Do you have any predictions? [116:55] For sure, I'd expect to see more of those. [116:57] Again, I would just say, look, what is the macro thing, [116:59] the big macro thing happening? [117:00] And I love what those guys are doing [117:02] and I love what Substack's doing, [117:04] but the big macro thing, [117:05] if you just think about the world change, [117:07] the big macro thing is TikTok, Instagram, [117:09] and then short-form video on X, [117:11] and a handful of other platforms. [117:13] That's the macro thing, and so, [117:16] where the future of the macro culture goes. [117:20] I mean, look, read Substacks, [117:22] but a thousand, or 10,000, [117:26] or a hundred thousand times more activity [117:27] is happening on TikTok, [117:29] and so the macro culture is going to be shaped, I think, [117:32] much more by short-form video, [117:33] at least for the foreseeable future, [117:35] and then, as I'm sure is obvious now, [117:38] but, the role of AI production [117:40] is about to really change things. [117:43] And there also may be the fact [117:46] that there's a single global feed now, [117:48] the fact that there's much less personalization in a way, [117:50] because so many things go to the top, [117:53] and in a way, I really actually don't like [117:57] the number of videos in my X feed these days. [117:59] I'm sure they perform in the metrics or something like that, [118:02] but if I wanted to scroll TikTok, I'd open TikTok, [118:04] and I don't want all the TikTok videos that get crammed in. [118:08] Do you guys get these in your feeds [118:10] where you get just random TikTok videos [118:12] from random accounts in your Twitter feed? [118:14] Yeah. Like, "No, [118:15] I'm reading a newspaper here. I'm not trying to watch TV." [118:17] Yeah, no, I think the people who run these things [118:20] have talked about this publicly, [118:21] but, yeah, all the old algorithms of [118:23] things that your friends like, [118:25] those are not as effective as [118:26] just the macro algorithm. [118:29] We are all more similar than we think. [118:30] Yeah, that, but but also the nuances [118:33] and interconnections are more subtle. [118:36] It's not the people you know, [118:37] it's the people you don't know [118:39] that you have connections with. [118:40] Yeah, exactly. Right. [118:41] You're probably more like a lot of other people [118:42] you've never met than you are like the people you know. [118:43] Yeah. For example, there's that. [118:46] By the way, having said that, [118:47] I believe the biggest, I think everything [118:49] we just talked about is very important. [118:50] I think the biggest, biggest, biggest thing that's happening [118:52] is just, we really, I think, for the first time, [118:54] we're entering the true era of free speech, [118:57] and I think that we started to get at that [119:00] in the 90s and 2000s, [119:02] and then there was a big reversion in the 2010s [119:04] with the censorship industrial complex that formed up, [119:07] and all the policies and all the government interference [119:09] and so forth, and, of course, a lot of that [119:11] continues on the part of the governments in particular, [119:14] but, in the US at least, that project has failed, [119:18] and the platforms themselves are really liberalizing out, [119:21] and then just the sheer volume, and scope, [119:24] and variety of content, and the number of ways [119:26] that people have to get messages out [119:28] in all kinds of ways, in the hyper-acceleration culture [119:32] where the sensors don't even know what to ban, [119:34] because they don't even know what half the stuff means. [119:36] We probably are living in the only true mass era [119:40] of free speech in human history. [119:44] You're seeing things now, [119:47] this is, although, probably not real time, [119:48] but you just see things now as just a normal user [119:50] that you never would've seen 10, [119:52] or 20, or 30, or 50 years ago. [119:53] There's not even a chance. [119:54] So, does this get to a political realignment? [119:55] I believe it does. Yeah. [119:57] So, I think this is the big thing, [119:58] I think Martin Gurri, [119:59] the guy who has published this book. [120:01] I think he really nailed it, [120:02] and I think his thesis and his book came out in 2015, [120:05] and I think a lot of people said either, [120:08] "Wow, he predicted Trump," [120:09] which is true to some extent, [120:11] but that's not the big thing that he predicted. [120:13] And then I think his prediction is, in some ways, [120:14] so fundamental that it's easy to just take it for granted [120:17] and say, "Oh, of course that's what's going to happen," [120:18] but it's actually so fundamentally important, [120:20] I can't stop thinking about it, [120:21] which is basically true transparency, [120:23] true transparency, true free speech. [120:27] It's just a fundamental solvent [120:29] basically dissolving all [120:30] centralized institutional authority. [120:32] And the reason for that is [120:34] centralized institutional authority is never perfect, [120:36] and it often has problems, [120:37] and, in fact, it often has very deep [120:39] and severe problems, as we were just discussing, [120:41] and the show that a government agency [120:45] or a big company could put on [120:46] to claim that they're better than they are, [120:47] that would've worked in under centralized media, [120:49] just simply collapses under conditions of [120:52] true free peer-to-peer communication. [120:54] There are just too many examples [120:56] of too many things that go wrong, for any institution, [120:58] for them to retain their credibility. [121:00] And then Martin and I have this big debate about this. [121:03] Martin and I have talked about it, [121:04] we have this big debate, which is, I'm like, [121:05] "Wow, that's fantastic," and he's like, [121:06] "No, Marc, I didn't mean this was good. [121:07] I never said this was good." [121:09] He said to me, the following. [121:09] He said, "Look, it is true that every major institution [121:12] is much, much more broken than they have been putting on." [121:15] He said, "However, it is also true that we do not know [121:17] how to run a society without large centralized institutions, [121:19] and so," he said, "those of you," like me, [121:21] "who cheerlead the collapse of centralized institutions [121:23] have not yet come up with an answer [121:25] for what exists on the other side." [121:26] But anyway, point being, I think [121:29] now, we're really going to go through that. [121:31] Now, we're really going to find out. [121:32] The business version of this is, [121:33] you used to be able to push a bad product [121:35] through a strong channel, [121:37] with strong marketing and sales. [121:38] Mm. Right. [121:39] You just can't do that anymore. [121:40] Right. The product quality will out [121:42] and it's deterministic. [121:44] Yeah, that's right. [121:45] And, by the way, you get this phenomenon. [121:47] You see this all over the place, [121:48] and you see this in Gallup does this great poll [121:50] of trust in institutions, and the numbers are just all [121:54] cratering, and the declines are accelerating. [121:58] You also see, and again, not to pick on specifics, [121:59] but you also see this in these political parties, [122:01] and you have a lot of this happening in Europe right now, [122:02] where these parties come in and they have, whatever, [122:04] 60% approval or whatever, [122:05] and then six months later, they have 15% approval. [122:09] It's like, "What the hell?" [122:11] Mm-hmm. Right? [122:12] Or, I'll give you an American example. [122:14] Eric Adams in New York, as the incumbent, has 9%. [122:18] It's just, how can you possibly have a system [122:22] in which the ruler has a 9%-approval rating? [122:25] Well, it's like, "Well, how did that happen?" [122:26] Well, it's all too transparent. [122:28] Everything that's going wrong is too transparent. [122:30] It can't be finessed. [122:33] The extreme version of this, [122:34] for good or ill, is that the centralized state [122:37] is an outcome of centralized media. [122:39] The nation state is downstream from the newspaper. [122:41] Right. Yeah, that's right. [122:43] Right, exactly, right, and so, yeah, [122:44] you can't hold it together. [122:48] You know, another of my counter arguments to Martin was, [122:53] basically, if you look at what the media landscape was like [122:55] in Colonial America, it was actually much more like [122:57] what it's like now than it was like it was in 1950. [123:00] Pamphleteers. Pamphleteers, and you'd have [123:02] 15 small newspapers in a city like Philadelphia, [123:04] and you'd have just enormous amounts of [123:07] contention and name calling, [123:08] and all kinds of things. [123:09] Anonymous bloggers. [123:10] Anonymous bloggers, yeah. They had all that stuff. [123:13] Benjamin Franklin literally wrote [123:14] under 15 different pseudonyms and he would [123:16] send them to people— [123:17] Voltaire, the same. Yeah. Fighting with each other. [123:18] Right, all these things. [123:20] And I was, "Look, we've lived this before," [123:22] and he's like, "Yes, and it was a time of revolution." [123:27] Correct, and to me it's so fascinating. [123:33] We're really in that now. [123:34] I feel like that was still being held back, [123:37] as late as last year, by the censorship apparatus, [123:39] and now, it's just like, "Okay." [123:41] Now it's all coming out. [123:43] And, maybe another way to think about this [123:45] is the narrative for the last decade [123:47] has been the internet is a fountain of misinformation, [123:49] and there is some truth to that. [123:50] There's a lot of misinformation online, [123:52] but the other thing is, [123:53] according to the Martin Gurri thesis, [123:54] the internet is an x-ray machine, [123:56] because every actually correct thing [123:58] that all of these institutions are doing wrong [124:00] is now being fully ventilated for the first time ever, [124:03] and they cannot survive that, [124:04] and that may ultimately include the governments themselves. [124:08] We're describing one trend here, [124:10] which is the move along [124:12] the decentralization centralization spectrum, [124:15] and, I think I'm not quite as enthusiastic, [124:17] or it seems pretty complex, that whole spectrum, [124:21] but the other change, to me, again, [124:23] seems to be the single global feed that's emerging. [124:25] So take an example, the Astronomer CEO, [124:28] and that whole thing, the CEO with his HR lady [124:31] being caught on video, that was just the front page [124:35] of the internet for that day, [124:36] or those one or two days. [124:37] I was talking to someone who was saying [124:38] they were talking to someone in China, [124:40] and they were joking about it, [124:41] but it was just prominent in China as well, [124:44] in the news there, and that didn't happen as much [124:49] 10 or 20 years ago, [124:50] and I don't even think it happened again, [124:52] even when we had the internet and cable media, [124:53] because we didn't have the clip [124:55] and the ability for the things to go as— [124:56] Yeah, I guess text is much more [124:58] language barrier. There's less virality. [124:59] But it's also language barriers [125:01] prevent text from crossing borders. [125:04] Clips can cross borders. [125:05] And just recommender algorithms [125:06] for things right at the top, I think. [125:08] So, all these factors, [125:09] do you just have thoughts on the implications [125:12] of having a single global feed? [125:14] Yeah, so this is kind of the global monoculture. [125:18] Maybe, I mean... Well, Marshall McLuhan [125:22] had this concept he called "The Global village," [125:24] and this is another one of these things where he said, [125:27] "People think I meant it positively and I actually didn't," [125:30] and so he said, "Electronic media formed the entire world [125:31] into a global village." [125:32] And he was talking about TV, [125:33] but you could say TV had a certain [125:35] amount of that too. Early version of that. [125:36] An early version of that, because it could spread [125:37] a single video feed much, much more broadly. [125:39] And what he said is like, "It used to be that every [125:44] village was its own village, [125:45] and so the things that happened in that village, [125:47] if the wrong man kissed the wrong woman, [125:48] it was a really big deal in that village, [125:50] but it wasn't a big deal in the next village. [125:51] Didn't even know about it. [125:52] Now all of a sudden, the entire world is becoming [125:55] a single global village." [125:55] And he said, "Here's the problem with that, [125:58] is that villages are really fucked up." [126:00] They're really dysfunctionally [126:02] fucked up a lot of the time, [126:03] because they're panopticons, everybody sees everybody else, [126:06] they're tremendously judgemental, [126:09] the social relations have carried tremendous weight. [126:11] If you end up getting sideways with the social relations [126:13] of the village, you're in serious trouble. [126:15] You might get exiled. You might die. [126:17] They're prone to manias and panics, witch trials, [126:20] They're hot-house environments. [126:23] They tend to go crazy. [126:25] And then specifically, I think, the next version of that, [126:29] I don't know if he said this, but other people said this, [126:30] is like, cosmopolitan societies [126:35] are like written written. [126:36] They're written in nature, [126:39] and they have the ability to have [126:41] dispassionate communication discussion. [126:43] Villages are all about orality. It's all oral. [126:47] It's all spoken, and, again, [126:51] it's this social hot house of spoken, [126:54] and therefore, de-intellectualized, [126:57] highly-emotionalized content. [126:58] So, Marshall McLuhan was writing about TV culture, [127:00] but he was actually pressing [127:01] into the clip culture. He was actually writing [127:02] about... I believe that's right, [127:04] and then, I think what he would say [127:05] if he were here today is, I think he would say, [127:07] "Yes, congratulations, guys. You got the global village." [127:09] He would say, "The Bible has the parable [127:11] of the Tower of Babel being a disaster [127:13] for a very specific reason. [127:15] If you centralize everybody [127:17] into a single, giant village, [127:18] you're going to have all the dysfunctionality, [127:20] you're going to have all the crazed panics [127:23] and freakouts of a village, [127:25] basically happening all the time," [127:26] which is, in fact, what we see. [127:29] You know, I think our friend Tyler Cohen, [127:35] at this point thinks this is all very bad. [127:38] McLuhan definitely thought it was bad. [127:39] On the other hand, I don't know, I grew up in a small town, [127:43] it wasn't that great. [127:44] A disconnected small town [127:45] wasn't that great either. [127:50] Is it really better to live in a world [127:51] where there's only a few places [127:52] where there's access to advanced thinking [127:55] and cosmopolitanism, or is it actually [127:57] the fact that everybody in the planet [127:58] can now be a full part of society and culture? [128:00] I feel like there's a Marc Andreessen worldview [128:02] that you've talked about enough [128:03] that it's now kind of a thing that exists beyond you, [128:06] that's maybe just being dispositionally optimistic [128:08] on technology generally, and refusing to break [128:13] any false nostalgia about the past. [128:16] Like, "I was there in rural, [128:18] small-town Wisconsin. [128:19] It wasn't good." You know? [128:20] Right, yeah. Exactly. [128:21] Yes, exactly. That's right. [128:22] That's right. I agree. [128:23] Great. Thank you. All good. [128:24] To cheeky pints. [128:25] Exactly. There you go.