Jeetu Patel 如何让 Cisco 兼具创业公司速度与大型企业规模

简要描述Brief Description

Jeetu Patel 如何将思科打造成全球最大初创公司般运作的企业How Jeetu Patel Runs Cisco Like the World’s Largest Startup

目录Table of Contents

• 思科押注人工智能
  Cisco’s Bet on AI
• 像初创公司一样运营思科
  Running Cisco Like a Startup
• AI 智能体所需的基础设施
  The Infrastructure AI Agents Need
• 让 32,000 名工程师拥抱 AI
  Getting 32,000 Engineers to Embrace AI
• 将 AI 引入硬件与芯片领域
  Bringing AI to Hardware and Silicon
• 深入了解 AI 数据中心建设热潮
  Inside the AI Data Center Buildout
• 我们是否正在过度建设 AI 基础设施？
  Are We Overbuilding AI Infrastructure?
• 建立对 AI 智能体的信任
  Building Trust in AI Agents
• 从印度到 Sizzler 餐厅当服务员
  From India to Waiting Tables at Sizzler
• 选择学习，而非维护自尊
  Choosing Learning Over Ego
• 雄心、团队合作与复利效应
  Ambition, Teamwork, and Compounding
• 不要浪费你的机会
  Don’t Squander Your Opportunity

思科押注人工智能Cisco’s Bet on AI

我的一个坚定判断是，展望未来，我们实际上需要更多工程师，而不是更少。目标不是提高效率，而是能够走出去，确保我们从 AI 生成的原创洞见中获益，从而解决过去连想都不敢想的问题。你知道吗，如果能回到 19 岁，50 多岁的人愿意花多少钱？身处当下，投身科技行业，真是个神奇的时代。今天我将与思科总裁兼首席产品官 G2 Patel 对话。我们将探讨 G2 如何把创始人思维带到思科、如何全力押注 AI、他们身处核心位置的这场大规模数据中心建设正在发生什么，最后还会聊聊创业者能从 G2 非凡的人生故事中学到什么。My strong assumption is moving forward that we will actually need more engineers, not less. The goal is not efficiency. The goal is to be able to go out and make sure that we benefit from the original insights that AI generates that allow us to solve problems we could have never dreamt of solving before. Do you know how many people in their 50s how much money they would pay to get back to being 19 if they could? This is a magical time to be alive in tech. Today I'm talking with G2 Patel, Cisco's president and chief product officer. We're going to discuss how G2 is bringing a founders mindset to Cisco, going allin on AI, what's happening inside the massive data center buildout that they're at the center of, and then we'll talk about what founders can learn from G2's incredible personal story.

G2，非常感谢你来到这里。G2, thanks so much for being here.

嘉宾：谢谢你邀请我。很高兴见到你。你的经历真是非同寻常——从在印度长大、在 Sizzler 餐厅当服务员，到后来成为世界上最大公司之一的总裁。不过，在聊这些之前，我想先聊聊今天的思科，以及你独特的职责和你在那里做的事情。我对自己职责的理解是，我要确保我们帮助打造出世界上最好的产品，让所有客户都能大规模使用。不过，思科所做的事情比我的工作更有意思：我们认为自己就像淘金热时期提供镐和铲子的公司。在当下这场大规模变革中，也就是 AI 时代，我们是一家 AI 基础设施公司。这是一种长期趋势性的转变。我们希望确保提供恰当级别的基础设施，让 AI 按照你期望的方式运行，同时确保运行安全、可靠。Guest: Thanks for having me. It's great to see you. you have this incredible background um from growing up in India, waiting tables at Sizzler to ending up as the president of one of the world's largest companies. But before we talk about all that um I want to talk about like Cisco today and your unique role and what you're doing there. The way I think about my role is my role is I just make sure that we help build the best products in the world that all of our customers can use at scale. What Cisco does though, which is even more interesting than what I do, is we think about us as the picks and shovels company during the gold rush. We are the AI infrastructure company or the AI era during this massive movement that's going on right now, which is a secular shift. And we want to make sure that we provide the right level of infrastructure so that it just AI just runs the way that you want it to and it runs securely and safely.

我很喜欢你对思科的这个描述——它是人工智能时代的“卖铲人”。显然，四年前你不会这么形容思科。所以这一切都发生在最近。这是一家规模极其庞大的公司所经历的转型，而它原本就已经在以惊人的规模运作了。嗯，在这样的时间跨度内加入这样一家企业，并尝试推动如此大规模的组织转型，感觉如何？And that description of Cisco, which I love, picks and shovels for the AI era, clearly you would not have described Cisco that way four years ago. So all of this is recent. This is a transformation of a very large company that's already was already operating enormous scale. Um, what has it been like to come in and try to transform an organization at this scale at that sort of time frame?

主持人：你知道，我觉得在公司发生这种大规模转变时，领导者通常会得到不该属于他们的过多赞誉，而真正干活的人却得不到足够的认可。但事实是，在我们做这件事的时候，我一行代码都没写，做这件事的是团队。我们做的是确保方向正确，在合适的岗位上聘用合适的人，并建立恰当的机制，弄清楚要开拓哪些市场，同时让每个人都觉得自己有机会赢。Host: You know, I think in these massive kind of shifts that happen in companies, uh, typically the leaders get an undeserved level of credit and the people that do the work don't get enough of the credit but the reality is is I didn't write a single line of code when we were doing this it is the team that did that u but what we did was we made sure that we actually set the right direction hired the right people in the right jobs made sure that we had the right kind of mechanisms to figure out what markets we're going to go pursue and do it with everyone feeling like they can they can win

嘉宾：你能讲讲你是怎么加入思科的，以及你现在在那里做什么吗？Guest: can you tell the story of how you ended up at Cisco and what you're doing there.

主持人：后来我去了思科。我当时在 Box，嗯，负责产品工作，和 Aaron Levy 一起共事。你知道，他自己也是一位创始人。所以我来到 Box 后，自己的整套思维模式又一次得到了强化，因为几年前我创办过自己的公司，之后去了 EMC。EMC 是一家非常大的公司。我当时心想：“啊，这里真是太企业化了。”后来我去了 Box，心想：“好吧，这里应该会很有意思，因为你真的可以像我以前经营自己公司那样来做事。”我觉得人分两种。一种人非常擅长担任职能部门的领导，也喜欢待在自己的职责范围内。他们会说：“别来告诉我在我的职责范围内该怎么做，我也不会告诉你在你的职责范围内该怎么做。”另一种人经营过某种业务，他们把自己视为业务的主人，而且完全不介意经常插手别人的业务。Host: So I ended up at Cisco. I was a box. I was uh running product with um you know for Aaron uh Levy um who is a founder himself and so my entire kind of mental model got reinforced again when I came to box because I had started my own company years ago and then I went to EMC. EMC was a very large company. I'm like ah this is like very corporate and I went to box. I'm like, "Okay, this can be fun because you can actually operate just like I used to operate in my own company." And I think there's two kinds of people. There's folks that are very good functional leaders and they love being in their lane. Uh, no one tell me what to do in my lane. I'm not going to tell you what to do in your lane. And then there's people who've run some kind of business and they think of themselves as owners of the business and they are completely comfortable poking in other people's businesses all the time.

而我And I

像初创公司一样运营思科Running Cisco Like a Startup

往往都属于那一类。所以我个人觉得，当你采用这种创始人模式时，它非常适合战时，也非常适合市场正在经历转变、事物变化节奏极快的时期。你要做的就是确保自己不会彻底疏远那些可能不处于创始人模式的人。tend to be of that ilk. And so I personally feel like when you have this founder mode approach, it's very good for wartime and it's very good for a time when the market's going through a shift, things are moving at a very fast clock speed. And what you have to do is you just have to make sure that you don't completely alienate the people who might not be in founder mode

嘉宾：并且在某种程度上帮助他们理解，如果你进入创始人模式，我们实际上就能以一种与迄今为止截然不同的节奏运作。这就是我们开始在思科做的事情。我觉得我会告诉我的同行和公司里的每一个人，我说，我觉得自己是思科一位后期加入的联合创始人。我晚了 36 年才加入，但我觉得这家公司是我的，这就是我的公司。如果有人从外部试图搞垮这家公司，那可得当心，因为我一定会确保自己为它而战。我们必须确保自己为工艺和公司的理念感到无比自豪。所有事情都是你的职责。所以，如果销售部门有什么事情做得不好，我就去和那边的人谈谈；如果市场营销部门有什么事情做得不好，我也会去和那边的人谈谈。Guest: and kind of help them understand the benefit that if you do go in founder mode, we can actually operate at a very different frequency than what we've been operating so far. And so that's what we started doing at Cisco. And I feel like I tell my my peers and everyone at my company, I'm like, I feel like I'm a latestage co-founder of Cisco. I was 36 years late coming in, but I feel like I own the company and this is my company. And if um if anyone from the outside tries to um screw with the company, beware because I'm going to make sure I fight for it. And we got to make sure that we we take a lot of pride in the craftsmanship and what the company stands for. Everything is your job. And so if something's not working well well in sales, I'll go talk to the people over there. Something's not working well in marketing, I'll go to talk to the people over there.

而且我完全料到他们会来挑我生意的毛病。And I fully expect them to come poke holes in my business.

主持人：这是不是你喜欢在思科引入前创始人的原因之一？你知道，我也意识到——嗯，也很欣赏经验丰富、具备系统思维的运营者的价值。Host: Is this one of the reasons why you like to bring in former founders at Cisco? You know, I also recognize the um and appreciate the value of seasoned operators who are systems thinkers

嘉宾：因为如果把思科所有当过创始人的人都请来，那节目可就有看头了。Guest: because if you had everyone at Cisco who was a founder, it'd be a show.

主持人：嗯。Host: Yeah.

嘉宾：嗯，所以你得确保达到良好的平衡。我们的做法是，嗯，我试着把它称为“三分法”。我的领导团队中有三分之一的人懂得如何在思科运作，都是能够在大规模环境下运营的人。第二个三分之一则是我从外部、从竞争对手和市场上引进的人，他们非常了解市场……Guest: Uh so you have to make sure that you get you get a good balance. The way we've done it is we have um I try to call it a rule of thirds. So a third of my leadership team are people that know how to maneuver Cisco that are at scale operators. The second third are people that I bring in from the outside from competition from the market that are very market in

主持人：嗯，他们可能不知道如何在思科内部运作，但他们是系统思考者，也曾在大规模环境中工作。然后第三类，第三类是我们收购的公司的联合创始人和 CEO。Host: uh that might not know how to maneuver Cisco but they're systems thinkers and they've operated at scale and then the third third are co-founders and CEOs of companies that we've acquired

嘉宾：但对于那些人，我们给他们的授权范围比他们卖给我们的公司还大。Guest: but those people we give them charters that are larger than the company that they've sold us.

主持人：你之前说过，这种运营方式在战时尤其重要。你觉得思科正处于战时状态吗？我觉得思科现在正处于战时状态。嗯，从某种意义上说，我们已经[哼笑]实现了产品层面上你见过的最成功的转型之一。嗯，我认为我们现在已经稳稳地进入了人工智能基础设施领域。这是三四年前没人会认可我们的地方。但展望未来三、四、五、七年，我们面前的机遇规模之大，远超以往经历的一切，所以我们必须确保能够抓住所有这些机遇。这意味着我们必须以……我喜欢在内部告诉团队，我们应该像全球最大的新创公司一样运营。这意味着既要快速行动，又要具备规模。如果你只是一味追求速度，那你就只是家新创公司。Host: You said earlier that this way of operating is especially important in wartime. Do you feel Cisco is in wartime? I feel like Cisco is in wartime right now. Um in the sense that we have [snorts] we have made one of the most successful turnarounds that you've seen from a product perspective. Um I think we are now squarely in the AI infrastructure game. Uh which is something that no one would give us credit for three years 3 four years ago. But if you look forward the next 3 4 5 7 years um the amount of u opportunity that exists in front of us is so massive compared to anything that we've experienced before that we have to make sure that we can we can capitalize on all of that and that means that we have to operate at I I like to tell my team internally we should operate like the world's largest startup that means you have to operate at speed but with scale if If you just operate at ski at speed, then you're just a startup.

如果你只是大规模运营，那你就只是规模最大的公司。如果你想成为世界上规模最大的初创公司，就必须兼具速度和规模。If you just operate at scale, then you're just the largest company. If you want to operate as a world's largest startup, it has to be with speed and scale.

基础设施 AI 智能体所需的条件The Infrastructure AI Agents Need

嘉宾：为了抓住这个机会，未来几年你们最需要把哪些关键事情做好？我们先假设，目前以高级用户的方式使用智能体的人不到2%，再想想如今生产出来的全部基础设施都被消耗殆尽这一事实。我们的基础设施短缺，供应受限。Guest: What are the key things you need to get right in the next few years to seize that opportunity? Let's assume right now that we have sub 2% of the of humans that are using agents in any kind of power user capacity and think about the fact that the entirety of infrastructure that gets produced today is getting consumed. We are short on infrastructure. We're supply constrained

主持人：只有不到 2% 的人使用智能体。Host: with only sub 2% of the people using agents.

嘉宾：想象一下，当 25%、30%、40%、50% 的人开始使用智能体时，你所需要的基础设施规模将会与我们现在大不相同。我们希望确保能够帮助解决这个问题。这是第一个问题。第二个问题是，这些智能体确实有时会失控。事实上，这周就很有意思，因为这些智能体有时会失控。它们甚至可能会有自己的想法。要区分智能体究竟是因为受到对手影响或被投毒而怀有恶意，还是只是认为自己做的是正确的事，只是在尽职尽责地执行你的目标指令，这件事正变得越来越难。Guest: Imagine when 25 30 40 50% of the people start using agents. the amount of infrastructure you're going to need is going to be at a very different scale proportion that we have right now. And we want to make sure that we can help solve that problem. That's problem number one. Problem number two is these agents, in fact, this this week was a very interesting week because these agents are going to go rogue at times. They might actually have a mind of their own. It's actually harder and harder to distinguish between when an agent has uh is operating with malicious intent because of uh an adversary influencing the agent or poisoning the agent or an agent just thinks it's the right thing to do and they're just kind of, you know, diligently going out and following your goal command.

主持人：paperclipip 优化器问题。Host: The paperclipip optimizer problem.

嘉宾：没错。是的。所以我们必须具备某种安全、防护和可观测性能力。这就是我们想要解决的第二个重大问题：确保每个人都能安全地使用 AI，不让人们因为持续的恐惧而不敢使用 AI。事实上，我认为情况恰恰相反。如果你不使用 AI，反而会更不安全。这是事实，因为大多数攻击都是以机器规模发生的。因此，你也必须具备机器规模的防御能力。我们是全球最大的网络基础设施提供商之一，也是全球最大的安全公司之一。我们希望确保做到这一点。Guest: Exactly. Yeah. And so what we have to have is you have to have some kind of safety, security, and observability capabilities in place. Um, and so that's the second big problem that we want to solve is we want to make sure that AI is safe to use for everyone. And that that's that doesn't become this constant fear because of which people don't use AI. Uh, in fact, I think it's the other way around. If you if you don't use AI, you will be less safe. And and that is in that is true because most of the attacks are happening at machine scale. So you have to have the defenses at machine scale as well. So we're one of the largest you know networking companies in the world that provide the infrastructure. We are also one of the largest security companies in the world. We want to make sure that we do that.

那么，我们也和 Splunk 一样，是全球最重要的机器数据公司之一。Then we are also one of the um you know one of the most significant kind of machine data companies in the world with Splunk.

主持人：哦，对了。所以这三者结合起来，应该就能为我们提供所需的所有底层基础设施，让我们成为 AI 时代关键基础设施企业所依赖的关键基础设施。Host: Oh that's right. And so the combination of those three should give us uh all of the underlying apparatus that's needed so that we can become critical infrastructure for the critical infrastructure companies for the AI era.

嘉宾：你在人工智能和数据中心市场进行了全方位布局，既投资这一市场，也将 AI 引入思科，让一切都变得更加高效。我真的很想了解，在思科这样规模的公司里采用 AI 是什么样的。这不能只是为了提高效率。目标不是提高效率，而是确保我们能够从 AI 产生的原创洞见中获益，从而解决过去连想都不敢想的问题。这和单纯让工作速度提高 20%，或者确保减少员工人数，是截然不同的。事实上，我坚信，未来我们需要的工程师会更多，而不是更少，因为每次我看到 AI 出现阶跃式提升时，都会随之出现一个瓶颈，而这个瓶颈恰恰在人身上。Guest: You've gone all in on AI both in terms of investing in the AI and data center market but also in terms of bringing AI into Cisco to make all make everything more efficient. I'm actually really interested to learn what it looks like to adopt AI at a company the size of Cisco. It can't be around efficiency. The goal is not efficiency. The goal is to be able to go out and make sure that we benefit from the original insights that AI generates that allow us to solve problems we could have never dreamt of solving before which is very different from just get to be 20% faster at doing this or make sure that you have less people. In fact, my strong assumption is moving forward that we will actually need more engineers, not less as we move forward because every single time I've seen a step function improvement in AI, I've seen a bottleneck that's emerged that's a human bottleneck.

你知道，编码实现自动化后，代码审查就会成为瓶颈。代码审查自动化后，判断该做什么、不该做什么就会成为瓶颈。所以，我确实觉得我们……我们需要……You know, coding is automated, code review becomes a bottleneck. Code review is automated. Judgment on what to build and what not to build becomes a bottleneck. So, I I do feel like we we're going to need to

让 32,000 名工程师拥抱 AIGetting 32,000 Engineers to Embrace AI

让更多人参与进来。但有趣的是，如何让 32,000 名工程师以及公司的其他员工：一是尝到甜头，二是不必担心使用 AI 会让自己陷入困境，或者可能丢掉工作。相反，我们采取了完全相反的做法：我们向员工保证，如果你不使用 AI，就会丢掉工作。have more people. But what what was interesting was getting 32,000 engineers and the rest of the company. One, to taste the honey, and two, not be fearful of the fact that if you use this, then you're going to actually be creating a trap for yourself or you might lose your job. Instead, what we did was we did something completely opposite. We guaranteed our employees that if you don't use AI, you will lose your job.

主持人：所以，这是一种非常不同的问题处理方式。这个说法并没有多么有争议，因为想想看，如果今天有人不会用电脑，也不会上网，你会雇他做知识型工作吗？绝不会。所以，如果你看到一种长期趋势性的转变，而且有一个庞大的平台即将借助 AI 进行重构，那么说不使用 AI 的人将不再那么重要，这并不是什么有争议的说法。因此，我们一直认为，公司和人只有两种：一种是非常熟练、灵活地运用 AI，另一种则会苦苦挣扎，难以保持竞争力。所以，我们向大家说得非常清楚：如果你恰好属于第二种人，因为你不用 AI，那么 Cisco 很可能就不适合你。Host: And so, it it was a very different way to kind of approach the problem. That wasn't a very controversial statement to make because think about today if you had someone that didn't know how to use a PC or didn't know how to use the internet like would you ever hire them in a knowledge worker job? You would never. So it's not that controversial to say that if you are seeing a secular shift and if there's this massive platform that's going to get refactored with AI then anyone not using it is just not going to be that relevant. And so we always believe that there's going to be only two kinds of companies and two kinds of people. Ones that are very fluent and dextrous with AI and the others that are going to be really struggling for relevance. And so we just made it very very clear for people that if you happen to be in that second camp because you're not using AI, chances are that Cisco is the wrong place for you.

因此，你们应该全力投入。我们会为你们提供所需的所有工具和支持，让你们全力投入。起初，我们实际上给了每个人无限量的 token，让大家直接开始使用，因为在我看来，这就像骑自行车：你得先熟悉一件事，然后逐渐掌握它，最后才能熟练高效地使用它。And therefore what you should do is go all in. And we will give you all the tooling and all the support to go all in. And initially we actually gave everyone unlimited tokens to just start using it because the way I think about it is you have to get it's like riding a bicycle. You have to first get familiar with something then you have to get good at it and then you can become efficient with it.

嘉宾：是的。Guest: Yeah.

主持人：但如果你一开始就追求效率，就永远不会熟悉，也永远不会变得擅长。所以我们开始这么做时，很多人一开始都强烈抵触。头几个月很艰难。对我来说，那个关键时刻是这样的：我团队里有一位工程负责人，是我见过最出色的人之一。他总是持怀疑态度，总会告诉我：“Ju，这些想法太疯狂了。我觉得永远不可能实现。”有一天晚上，他给我打电话，态度非常诚恳。他对我说：“嘿，我想告诉你，我错了，你是对的。现在模型的发展速度太快了。”我也受益于一些信息，因为你知道，Kevin Weil 在我们的董事会里，而且我也认识生态圈里的一些人。每次我和创业圈的人交谈，得到的信息都和我在内部了解到的截然不同。Host: But if you try to go efficiency first you'll never get familiar or you'll never get good. And so we started doing that got a lot of initial resistance from a lot of folks. The first few months were rough. And the the seinal moment for me was I have this one engineer who's engineering leader in my team, one of the most amazing people. And he is always a skeptic. He's always going to tell me, Ju, those ideas are crazy. I don't think it's going to ever happen. And he called me one day in the evening and he was as sincere as he could be. And he said to me, hey, I just want to let you know I was wrong. You were right. The models are moving at a pace now. And I had some benefit because you know Kevin wheels on our board and I had some you know kind of I I knew some people within the ecosystem that every time I talked to someone in a startup world um I was getting a very different set of data than I was I was I was getting internally.

于是，威尔最后对我说，事情现在已经进展到这一步了：我们把大概五十万行代码重构了一番，最后精简到了大约十二万行左右。And so Will says to me finally this has now moved where we have now taken and refactored like you know half a million lines of code and we were able to refactor it into like 120,000 lines of code or something

嘉宾：因此性能提升了 5 倍，而我们大概在两周内就做到了这一点。Guest: and so 5x improvement and we were able to do that within like 2 weeks

主持人：你知道，所以你说得对，这件事确实会进展得非常快。Host: you know and so one you are right that this is actually going to move really fast

嘉宾：所以我不能只考虑我们今天拥有什么，还得想想三个月后会发展到什么程度，不是一年后，而是三个月后。要以此为前提，从今天开始做出改变。Guest: so I can't just think about what we have today I have to think about the fact of where is this going going to be in 3 months from now, not a year from now, 3 months from now. And assume that and start making the changes today.

主持人：但第二点是，我也对这可能意味着很多人会面临许多糟糕的事情感到非常焦虑，而且我们可能不得不对员工进行很多调整。Host: But second is I'm also feeling very anxious about the fact that this might mean a lot of bad things for people and we might have to make a lot of changes in our staff.

嘉宾：然后我就想，或者我们其实可以激励每一个人。Guest: And I'm like or we can actually motivate everyone

主持人：确保我们都能因为使用它而有所提升。后来你知道，这种效果就不断累积。我们实际上成了 OpenAI 和 Codex 的第一家设计合作伙伴。他们给了我们一些 FTEEs。直到现在，我大概每两三周就会和整个工程领导团队坐下来，还有他们的 FDES，一起讨论过去两周哪些事情进展顺利，哪些出了问题。大家很清楚，我绝不会轻易放弃这件事。我认为，在大公司里，有些事情必须自上而下推动。大多数事情自上而下推动会相当困难，但有少数你视为信仰的事情，就必须自上而下推动，而且要确保自己坚持到底，毫不动摇。Host: to make sure that we can um we can all get better as a result of the use of this. Then you know it just kept compounding. We actually became the first design partner with uh with OpenAI and Codex. They gave us some FTEEs and then I to this date um every um I think it's every two or three weeks now uh I sit down with my entire leadership team of engineering and then their FDES and we talk about like what went right, what went wrong in the past two weeks and um it it was just very clear to people that I'm just not going to let this one go. And I think that's in large companies, there's a few things where you have to go top down. Most things it's going to be pretty hard to go top down, but a few things that you think are religion, you have to go top down and you have to just make sure that you're unrelenting on those.

人工智能就是其中之一。And AI was one of those.

将 AI 引入硬件与芯片领域Bringing AI to Hardware and Silicon

嘉宾：我们都见证了 AI 如何变革软件开发，但思科在硬件开发方面也投入了大量工作。Guest: We've all seen how AI is transformative for software development, but Cisco does a lot of hardware development.

主持人：YHost: Y

嘉宾：你们在硬件开发方面看到的影响也一样吗？还是说这仍然是未来才会发生的事？思科涉及软件、硬件和芯片，而这三者各自的采用周期都不一样。所以在硬件方面，你现在开始看到一些不错的设计模式逐渐显现出来。但这还不如软件领域那么明显。软件的发展速度非常快。至于芯片，团队有一定程度的审慎，不会贸然全面投入，因为他们非常在意确保数据和知识产权……Guest: are you seeing the same impact on the hardware development side or is that still a in the in the future? So the Cisco does software, hardware, and silicon. And each one of them have a different kind of adoption cycle. So on the hardware side, you're starting to see um good kind of design patterns that are starting to emerge now. U but it's not quite as much as software. Software at the clock speed is really fast. silicon there's a level of intentionality of not going allin by the team because they're very concerned about making sure that the data and the IP

主持人：不会离开，然后转眼间就有人复制了它。Host: does not leave and then before you know it someone else has replicated it

嘉宾：对。Guest: right

主持人：你知道，诸如此类。在硅片方面，我们之前的步伐稍微慢了一点，但现在，嗯，我认为随着模型越来越好、开放权重模型的出现，以及我们拥有自己的基础设施等等，情况会发生改变。不过，我们必须压缩硅片领域的周期时间。我们业务中有意思的一点是，我们是全栈企业，对吧？我们既制造硅片，也制造光子器件，还打造系统、硬件、软件，涵盖整个技术栈。技术栈中的每个组成部分和层级都有不同的规划周期。硅片的规划周期是五年。也就是说，我今天要做的，是确保自己知道五年后大规模生产会是什么情况，然后现在就做出相应的投资。Host: you know and so on the silicon side we had gone a little bit slower but now um I think with the models getting better um with open weights models being there having our own infrastructure all of that I think it it's going to change but we we have to compress rest the cycle time for silicon. The thing that's interesting in our business is we are a full stack, right? Like we build silicon, we build photonics, we build systems, hardware, software, the entire stack. Each one of the elements and layers of the stack have a different planning cycle. So silicon is a 5year planning cycle. What I do today uh I have to make sure that I know what's going to happen for atcale production 5 years from now and then make the appropriate investments today.

嘉宾：这也太离谱了。你怎么可能在 AI 领域做到这一点？AI 才三、三年大。想想整个制造周期吧，不只是芯片本身，还包括从芯片到系统、到 SDK，再到推向市场。你得确保早点启动，因为要制造芯片，就必须先构建芯片所需的 IP，比如 SerDes 以及所有这些组件，而这些实际上都得提前开发。所以，芯片领域需要按一定顺序推进的工作，不可能在短短一段时间内完成。你不能说“去造一颗芯片，12 个月后就能造出来”，虽然确实能造出来，但品质不会太好。硬件通常需要 18 到 24 个月的规划周期，有些情况下甚至要 30 个月，因为供应链的缘故。想想看，我们要大规模生产非常多的产品。Guest: That's bananas. How can you possibly do that in the AI? AI is only three three years old. think about the entire um manufacturing cycle of not just the silicon but silicon to systems to SDKs to taking it out to the market. You have to make sure that you start because in order to build the silicon you have to build the IP for the silicon like certis and all those pieces and those take actually um you have to build them up ahead of time. So like there's a sequencing that needs to happen in silicon that cannot happen within a matter of like you know you can't just say go build a chip in 12 months and you can build a chip it just won't be that good. Hardware tends to be um you know 18 to 24 month planning cycle maybe 30 months in some cases because of supply chain because we think about it we're doing this at scale for very large volumes of um of units.

你进入软件和操作系统领域，那是一个为期12个月的规划周期，也可能是18个月。然后你进入模型、智能体和应用领域，这时规划周期就变成了三个月，甚至可能是每周。我们必须确保这个协同设计的全栈真正让所有这些部分和谐运作。如果你们公司能做到这一点，就会拥有一道极具战略意义的护城河，因为你们已经找到了让这些不同层面协同运作的方式，并形成了自己的知识产权。这对其他人来说非常难以做到，也很难复制。You go to software and operating system. That is a 12 month planning cycle, maybe 18month planning cycle. Then you go into models and agents and apps. Now you're into a three-month planning cycle, maybe a weekly planning cycle. And we have to make sure that that codeesign full stack actually has all of those things working in harmony. And if you can figure that out as a company, you have a very strategic mode in the sense that you've been able to figure out all these different layers that work well together from an IP. That's that's very hard for other people to do and replicate.

主持人：现在有了 AI，能否加快这些周期？Host: Is it possible to speed up those cycles now that AI is here?

嘉宾：当然。它们都会被压缩。但我说的规划周期，并不只是指建造它需要多长时间。我、我是说，从构想到大规模生产并创造数十亿美元营收的整个过程。说到芯片……Guest: Absolutely. All of them will get compressed. But when I say planning cycle, I don't just mean how long does it take to build it. I I mean from idea to at scale production with billions of dollars in revenue. Speaking of chips

深入 AI 数据中心建设现场Inside the AI Data Center Buildout

以及硬件和整个供应链。我想谈谈数据中心的建设，因为正在发生一件非同寻常的历史性事件。我们正在建设的规模，超过了一代人以来的任何时候，而你正处于这一切的中心。能谈谈思科在这场大规模数据中心建设中扮演的角色吗？ 顺便说一下，在数据中心建设中，每个人都知道 GPU 的作用。当时，刚开始出现这些 Transformer 模型时，它们可以放在单个 GPU 上运行，所以训练过程相当简单。后来，参数越来越多，GPU 也需要扩展，于是一台服务器里装了八个 GPU。这样一来，计算就扩展到了服务器层面，但这些 GPU 仍然都在一台服务器内互连。接着出现了服务器机架，GPU 通过一台阿罗科服务器互连，协同工作，就像一台巨型机器。随后，规模又从机架扩展到了机架排，因为大型数据中心里现在有多排机架需要协同工作。and hardware and this whole supply chain, I want to talk about the data center buildout because there's this incredible historic event happening. We're building more stuff than we built in a in a in a a generation and you're at the center of it. Can you talk about the role that Cisco is playing in this huge data center buildout? So in data center buildouts by the way everyone knows the role the GPU plays and what had happened was when you just started seeing these transformers come about these transformers would fit on a single GPU and so the the training runs would be pretty simple and then uh you had more parameters and those GPUs needed to scale and so you had eight GPUs within a server and so you now went into a server but those GPUs were all interconnected within a server. Then you had racks of servers, GPUs got interconnected with an arocco server working coherently like one, you know, kind of giant machine with um things going and then racks started scaling to rows because you now had multiple rows of racks that needed to be worked together within a large data center.

嗯，这就需要他们所说的横向扩展网络，而不仅仅是为了机架级扩展而进行纵向扩展网络。现在，由于电力受到巨大限制且供应短缺，你无法总是把所有电力都输送到数据中心所在的地方。所以你要做的就是去电力供应充足的地方。因此，你可能会有相距数百公里的数据中心，而它们需要协同工作，成为一个完整的逻辑单元。Um and that required what they call scale out networking, not just scale up networking for um um for for rack level scaling. And now what you have is because there's such a huge power constraint and power shortage, you can't always bring all the power to where the data center is. So what you do is you go to wherever the power is available. And so you might have data centers that are hundreds of kilometers apart that need to then work like one completely um logical unit.

主持人：哦，两个相距数百英里的独立物理数据中心，需要协同工作，就像它们位于同一栋楼里一样。Host: Oh, two different physical data centers separated by hundreds of miles that need to beworked together as if they were in the same building.

嘉宾：就好像它们在同一栋楼里，就好像它们是同一台电脑一样。Guest: As as if they were in the same building as if they were the same computer.

主持人：哇。Host: Wow.

嘉宾：这就需要一种完全不同的网络架构，称为跨规模网络。在数据中心内部，你可能会采用横向扩展网络；在机架内部，则采用纵向扩展网络。我们的重点主要是横向扩展和跨规模网络。我们提供的不只是用于连接 GPU 的芯片和交换机托盘，还包括光子技术，也就是光学器件。两年前，我们开始向超大规模云服务商销售产品，用于 AI 工作负载。当时我们向市场提供了为期三年的业绩指引，并表示接单额将达到 10 亿美元。Guest: That requires a whole different family of networking called scale across networking. For within a data center you might have scale out networking and for within a rack you have scale up networking. We are largely focused on scale out and scale across and we do not just the silicon and the switch trays for connecting the GPUs together but we also do the photonics which is the optics. We started our business selling to hyperscalers two years ago for AI workloads and we thought we'd do we'd give the market uh this tree of guidance and said we'll do a billion dollars in orders taken.

主持人：我们第一年最终达到了23亿。Host: We ended up at 2.3 billion in year one.

嘉宾：所以第二年我们就说，那就翻一倍吧，不做23亿了，做个50亿左右。去年，也就是一个月前刚结束的第四季度，我们最终做到了90亿。Guest: So the next year we said well let's double it instead of 2.3 let's do like 5 billion. Last year which was a Q4 that just ended month ago uh we ended up at 9 billion.

主持人：哇。你这两年都实现了非常乐观的预测目标的两倍。Host: Wow. Both years you doubled the very optimistic projections that you made.

嘉宾：是的。所以我们基本上从零做到了93亿。Guest: Yeah. And so we basically went from zero to 9.3 billion

主机：十亿Host: billion

嘉宾：短短两年间。Guest: in a matter of two years.

主持人：哇。Host: Wow.

嘉宾：没错。所以这就是事情发展的速度。我们仍有很大的市场份额有待开拓和获取。这是一门复杂的生意，营收达630亿美元，产品有数百种，嗯，我们必须确保以正确的方式进行布局。Guest: Right. And so that's the speed at which this is moving. We still have so much share to go out and capture. It's a complicated business. $63 billion in revenue, hundreds of products um that we have to kind of make sure that we maneuver in the right way.

主持人：但在总计630亿美元中，这90亿美元对一条全新的业务线来说占比相当可观。90亿美元的订单要过一段时间才能转化为收入，不过……Host: But the 9 billion out of the 63 billion total is like quite material for a brand new business line. nine billion in orders. That takes a while to get to revenue, but

客人：我明白了。嗯。知道了。Guest: I see. Yeah. Got it.

主持人：但是，但是你说得对。这，这，这开始变得有意义了，这令人兴奋，因为你说得对，你正身处其中。Host: But but you're right. It's it's it's starting to get to be meaningful and it's exciting because it's you're right in the thick of things

嘉宾：而且这个市场非常庞大。一吉瓦的容量，建造成本过去大约是350亿美元，现在接近500亿美元。建设一吉瓦的数据中心容量，大概就是这个成本。其中约有10%到15%用于网络设备。成本主要在GPU上，但这里面有大量复杂的工作：你知道，你得建好建筑外壳，还得办理电力方面的许可。你走进一座数据中心，就会意识到这一切有多复杂。比如，你在提示词里提出一个问题，说“给我一个披萨食谱”，幕后要发生什么，才能在一两秒内把答案交给你？Guest: and the market is so huge. A g a a gigawatt of capacity would cost about anywhere between it used to be 35 billion now it's like close to $50 billion uh is what a gigawatt of data center capacity would would cost to build. And about 10 to 15% of that tends to be networking. Largely it's in the GPU and but there's a huge amount of complexity in you know you have to build out the shell and you have to build out the power permits for that and the you walk into a data center it impresses upon you on how complicated when you make a query on your prompt to say give me a pizza recipe. What happens behind the scenes to get that to you within um a second or two?

我们是否正在过度建设 AI 基础设施？Are We Overbuilding AI Infrastructure?

主持人：因为建设规模实在太大了。有人想知道，我们是不是建得过多了。我们是不是……Host: Because the buildout is so extreme. Some people have wondered whether we are overbuilding. Are we

嘉宾：绝对不是？让我告诉你原因。人们试图把这两者等同起来。首先，在人工智能领域不存在先验。所以，我从根本上确实相信这一点，而人们犯的一个重大错误就是总是回到自己过去的经历。我认为，在人工智能领域最重要的技能之一，嗯，就是学会忘却，因为过去的经验可能会成为巨大的负担。但人们通常会试图把这和互联网泡沫时期等同起来。天哪，当时发生了大规模的基础设施建设。当时的情况是，你在建设基础设施，并希望需求能够赶上供给。而现在，不管你建设多少基础设施，Guest: categorically not? And let me tell you why. And people try to equate this. Firstly, in AI there are no priors. So I do fundamentally believe that and this is one of the big mistakes people make is they keep going back to an experience they've had. I think one of the most important skills to learn um in AI is unlearning because your past experience can be a huge liability. But people typically will try to equate this to the dotcom boom days. Oh my goodness, there was so much infrastructure buildout that happened. What was happening during that time is you were building out infrastructure and you were hoping that demand would catch up to the supply. Right now, no matter how much infrastructure you're building,

主持人：它几乎瞬间就被消耗掉了。所以不存在什么供给不受限制的时期——尽管增长呈指数级，看起来几乎是一条垂直线。从这个角度来看，我确实觉得，嗯，不存在什么严重的过度建设。事实上，我认为情况恰恰相反：如果使用智能体的人还不到 2%，而这些智能体平均消耗的网络带宽是人类的 450%——我们执行的是完全相同的任务。现在想想数万亿个智能体……Host: it's getting consumed almost instantly. So it's there's no patch of time like you're supply constrained despite the exponential growth that looks like a vertical line. So I do feel like from that perspective I don't think that there's a um there's a huge um overbuild. In fact I think it's the other way around if sub 2% of the people are using agents and these agents are so consumptive on average an agent is 450% more consumptive on network bandwidth than a human would be. We're conducting the same exact task. Now think about trillions of agents,

主持人：对吧？Host: right?

嘉宾：嗯，现在想想，你的大多数任务都由智能体来执行，而它们必须确保不断补充上下文。你有一个技能文件、一个 Markdown 文件，还有 Gary 的 GStack，你知道，这个过程会不断来回进行。也就是说，会有很多轮来回沟通。Guest: Um, and now think about most of your tasks getting conducted by agents where they're going to have to make sure that they keep enriching context. You have a skills file, you have a markdown file, you know, you've got the GStack from Gary, and it's going to keep going back and forth. Like, there's going to be a lot of back and forth that happens.

主持人：这会消耗很多令牌。Host: It's going to be a lot of tokens.

嘉宾：是啊。说到智能体，我听说你最近开始喜欢上使用自己的个人智能体了。我很好奇，你使用智能体的情况是什么样的？我用它有一段时间了，但 Instinct 对我来说是个倍增器。这也就是大概 4 天前的事。我特别不擅长处理邮件。真的特别不擅长。大概 5 天前，我有 171,762 封邮件之类的，其中有 146,000 封都没读。Guest: Yeah. Speaking of agents, I heard that you've recently become a fan of using your own personal agents. I'm curious, what what is what does your usage of agents look like? So, I've been using it for a while, but instinct was the force multiplier for me. And this is just like 4 days ago. I I suck at email. Really suck at email. Like I I had 171,762 emails or something just like 5 days ago. 146,000 of them were unread.

主持人：好的。Host: Okay.

嘉宾：我凭直觉行事。我鼓起勇气，说：“你知道吗？我要把我的个人邮箱权限给它。”呃，我有200封邮件。哇。是啊。Guest: I used instinct. I took the leap of faith and said, "You know what? I'm going to go give it my personal email access." Uh I have 200 emails. Wow. Yeah.

主持人：它把你的整个收件箱都轰炸了一遍。Host: It blasted through your entire inbox.

嘉宾：它把一切都处理好了。它告诉我该做什么。它真的会把邮件移到合适的位置，并妥善归档。如果我需要什么，它就会把它找回来。它还会把我从所有来源收到的新闻新闻都总结给我。我、我当时觉得，虽然我其实不怎么处理邮件，但我潜意识里有一部分还是觉得不堪重负。现在我感觉好多了，哇，我现在能掌控这一切了。突然之间，我开始查邮件。我就想，我为什么要查邮件？嗯，因为只有三封邮件时，查邮件还挺有意思的。Guest: It took care of everything. It told me what to do. It actually moved things and archived them in the right way. If I needed something, it would bring it back. It actually summarized my news news for me from all the sources that I was getting. I I was feeling even though I didn't really do email that much, I there was a part of me subconsciously that was feeling overwhelmed. And I just feel better that wow, I'm I'm on top of that now. All of a sudden, I'm checking email. I'm like, why am I checking email? M because it's fun to check email when you have three emails.

主持人：这确实又回到了那个观点：数据中心的建设永远都不够。想想看，要处理 14 万封邮件并弄清楚该怎么做，需要多少 token。我觉得，某种程度上，关于“天哪，这就是个泡沫，一切都要完了”这类说法有些夸大，我并不觉得这有什么不幸。两种情况仍然可能同时成立：这会带来一场长期的转变，从根本上彻底改变人们的生活方式，而这需要大量基础设施；同时，也会有一些公司的估值非常虚高，因为每逢这类大规模的平台转型，都会出现大量尝试，最终只有少数能胜出，不会所有尝试都成功。所以估值会有些虚高，但在我看来，这并不构成泡沫。Host: It really ties back to the the point about the data center buildout never being sufficient cuz like think about the tokens to go through 140,000 emails and figure out what to do with it. This is where I feel like I just don't think it is unfortunate in some way that the level of kind of um exaggeration that happens on oh my goodness this is a bubble everything's going like you might still have the two conditions can hold true. You will have a secular shift where this is going to completely fundamentally change the way that people live and that's going to require a lot of infrastructure and you will also have some companies that are very frothy in valuation because from time to any any kind of massive platform shift like this you will have a ton of experimentation that goes on and only a few of those are going to win not not all of them win and so there will be some frothiness in the valuation but that does not constitute in my mind a bubble.

泡沫的问题在于，如果这种长期趋势本身受到质疑，我认为很难站出来为此辩护。bubble is if that secular shift itself is questioned and I think it's very hard to go out and argue for that.

建立对 AI 智能体的信任Building Trust in AI Agents

嘉宾：所以，嗯，最近新闻里一直在报道，嗯，这些 OpenAI 智能体失控了，入侵了 Hugging Face 以及其他一些地方。你怎么看现在 AI 安全方面发生的这些事？另外，这对在大型、任务关键型企业中部署 AI 有什么影响？Guest: So um it's been in the news recently this um open AI agents went rogue compromised hugging face and and other places. What are what are your thoughts on what's happening with AI safety now? And also how does that play into like deploying AI at a large like missionritical enterprise?

主持人：是的，我觉得这是个很好的问题。我现在想说的是，未来几年，这会成为我们这个时代最具决定性的话题之一。此前，算力受限和基础设施不足是个大问题，我们也一直在讨论是否有足够的算力——这仍然是个问题。但更大的问题在于，我能不能信任这些系统，把我所有的工作和个人生活都托付给它们。可信赖的委托与不可信的委托之间的差别，就是市场领先与彻底破产之间的差别。Host: Yeah, I think it's a great question. One of the things that I would say right now is it is going to be the defining um you know kind of topic of our time right now in the next um next few years because the constraint and infrastructure was a really big deal and we were talking about a lot about is there compute available that's still a problem but a much much larger problem is whether or not I can trust these systems that I'm going to delegate all my work to and all my personal life to. The difference between trusted delegation and untrusted delegation is the difference between market leadership and complete abject bankruptcy

嘉宾：在我看来，我认为我们必须确保真正解决这个问题，而这需要做很多事情。思科目前正在大力投入，确保能够提供一套完整的智能体安全平台，并将其推向市场。它涵盖方方面面：从全面了解哪些内容进入了模型，到管理非人类身份的身份信息、盘点环境中运行着哪些智能体，再到验证模型本身，其中包括算法红队测试，用来检验能否越狱这个模型；如果模型真的被越狱了，你问了它一个不希望它回答的问题，或者让智能体执行了你不希望它执行的操作，我们要如何拦截？又如何确保在运行时动态地设置并强制执行防护措施？我们提供整套能力，既包括安全防护，也包括可观测性，并将两者融合在一起。我认为，这是我们这个时代最重要的议题之一。Guest: in my mind. So like I think we have to make sure that we actually get through that and that requires uh a bunch of things and we are actually Cisco is working a lot on making sure that we can provide a whole agentic security platform that we'll we'll we are making available to the market. Uh that everything from having full visibility on what goes into a model to identity for non-human identities and inventorying what agents are running within your environment to actually doing validation on the models itself um which includes algorithmic red teaming to say can you uh jailbreak this model and if you do jailbreak it and you asked it a question that you didn't want it to um have it have a response or if you had the agent conduct an action that you didn't want it to conduct, how can you intercept that and make sure dynamically at runtime you can put you can enforce guardrails that entire apparatus we actually make available including the security side but also the observability side um and we fuse it together and I think that's it's one of the most important issues of our times.

主持人：很酷的是，你们不只是要在内部开发它，还打算把它打造成一款产品。Host: It's very cool that you're building it not just internally but you're actually going to build it as a product.

嘉宾：不，我们已经在市场上提供这项服务了。事实上，我们有一款名为 AI 防御的产品，它是这项服务的核心基础，一年半前就已推出。它的表现非常出色，我们只需要确保让越来越多的人用上它。最初，这款产品只针对聊天时代而打造。我们先将它应用于模型，随后又扩展到了智能体。现在，我们可以随时对智能体进行动态监控。你知道，如果你向智能体申请退款，它不应该退给你鞋子价格三倍的金额；而这种情况在零售业时有发生。我们收购了一家名为 Galileo 的公司，来自 DeepMind 和 Uber 的一些人也加入了进来。他们做得非常出色。Guest: No, we have it available in the market. In fact, we have a product called AI defense which is the core foundation of that that we launched a year and a half ago. It's doing ex exceptionally well and um we just need to make sure that we get to more and more of it and that was only done for the chat era. So we did that initially for the model and then we extended that to agents and so now what we have is dynamic monitoring of agents at all times and you know if you ask an agent for a refund it shouldn't be giving you three times the amount that you paid for the for the pair of shoes you know and that's something that happens all the time in in the retail industry and so that's we acquired a company called Galileo um there were folks from Deep Mind and um and Uber that um that came along And then those folks have done a great job.

呃，他们是 Splunk 团队的一员，而且这件事做得非常出色。Uh they're part of the Splunk team and they've done a great job on it.

主持人：能否带我们回顾一下你的职业生涯早期，帮助我们了解你成长的环境和早年从事的工作是如何让你走到如今这个独特的位置，并产生如此深远的变革性影响的？现在……Host: Can you take us back to your early career and help us understand how growing up where you did and the early jobs that you worked like led you to be in this unique position to have such a transformative impact. Now

从印度到 Sizzler 餐厅当服务员From India to Waiting Tables at Sizzler

嘉宾：现在回想起来，我当时非常幸运。那时候我并不觉得自己有多幸运，但我爸是个专门实施大额诈骗的骗子。Guest: when I look back now, I was very fortunate. At the time, it didn't seem like I was very fortunate, but my dad was a high stakes con man.

主持人：他就像伯尼·麦道夫一样，呃，会骗别人的钱。呃，有段时间我在印度生活变得很危险。嗯，有人想绑架我。他有个情妇，那情妇自杀了。我不想再待下去了。Host: He was like Bernie Maidoff. uh he would swindle people for money. Uh at some point it got dangerous for me to live in India. Um and there were people who were looking to kidnap me. He had a mistress. That mistress had committed suicide. I didn't want to stick around.

嘉宾：我的天啊。Guest: Oh my goodness.

主持人：你知道，我人生中的那段经历，完全可以拍成一部电影。所以我决定这么做。我父亲对我母亲非常暴虐，而我和母亲关系很亲近。于是我们俩都逃了出来。她决定，嗯，你知道，住到一个不对外公开的地方。我就说，让我先出去看看美国，看看能不能了解一下学校之类的。于是我就来到了这边。Host: You know, it was it was you could make a movie out of that that that that part of my life. And so what I decided to do and my he was very abusive to my mother and I was very close to my mother. And so both of us escaped and she decided to um you know uh live in um a place that was undisclosed and I said let me just go out and check out America and see if I can check out some schools and all of that. Came over here.

主持人：你多大了？Host: How old are you?

嘉宾：那时候我19岁。是1991年，对吧？所以，呃，我来到这里，叔叔让我暂住了一阵子。我在一家叫 Sizzler 的高级餐厅当服务员。嗯，平均每次小费一美元。你知道，那段时间发生了几件很关键的事，让我在心态上有了很大的转变，因为我在学校时成绩很差。嗯，我也不是那种……你知道，我算是家里的异类，因为我的很多表亲在学业上都很聪明，而我不是，所以我总有一种自卑感。我以前会结巴，所以总是保持沉默。嗯，我接电话时连“你好”都说不出来，所以当时有各种各样这样的问题。后来我开始当服务员，我就想，得了，你必须想办法逗客人开心，不然就赚不到钱。然后我来了，嗯，你知道，我在这边读本科时，接触到了一家叫 Docu 的公司。他们想给我一个实习机会，因为我的教授对我说：“嘿，你可以考虑去那里实习。”她是我在学校里的导师，人特别好，所以她，嗯，帮我引荐了。于是我就得到了一份实习工作。Guest: I was 19 at the time. 1991, right? And so uh I came here, my uncle put me up for a while, waited on tables at a fine dining establishment called Sizzler. um average tip of a dollar and um you know there were a few seinal things that happened during that time that really gave me a mental reframing because I was a terrible student in school. Um I wasn't really that you know I was a black sheep of the family in the sense that a lot of my cousins were all very smart academically and I was not and so I always had this kind of inferiority complex. I used to stutter and so I would always stay quiet. um I couldn't say hello when I picked up the phone and so there was all of these kind of things that were happening and then I started waiting on tables and I'm like boom you have to figure out a way to entertain or you will not make money and then I came um you know when I was over here I was in undergrad I came across this company docu and they wanted to give me an internship because my professor said to me hey you might want to get an internship here she was my mentor in school and she was great and so she um made an introduction And so I I got a job as an intern.

大概几个月后，那个人——创始人——来找我，说：“嗯，我们原来的投资人要退出了，我们打算把他们的股份买下来。”我说：“哦，我也想参与。”他们说：“嗯，我们每个人都要投 25 万美元。你也可以投 25 万美元，这样就能加入了。”我去找了一个我认识的人，他是个银行家。我问他：“我需要 25 万美元。”他说：“你为什么需要 25 万美元？”我说，我要买下这家公司。他问：“你买下公司之后打算做什么？”我说：“我要经营它。”要记住，我不是这家公司的创始人，我只是个实习生。Like few months later, the guy comes to me, the founder, and says, "Well, we're going to my original investor is going to um leave. We're going to buy them out." I'm like, "Oh, I want in." And they're like, "Well, each one of us put putting in a quart million dollars. You can put in a quart million dollars and you can have in." I went to one of my the people I knew and uh he was um he was a banker and I asked him like, "I need a quart million dollars." And he said, "Why do you need a quarter million dollar?" So I'm like, I'm going to buy this company. He goes, "What are you going to do when you buy the company?" I'm going to run it. Now remember, I wasn't a founder of this company. I was an intern.

主持人：你当时还是一名实习生，还在上大学。Host: You were you were an intern. You were still in college at the time.

嘉宾：我刚刚……我借的第一笔贷款是10,000。第二笔贷款是250,000，不过我借第一笔贷款时还在上大学。我拿的是最低工资。Guest: I had just The first loan I took was 10,000. The second loan I took was 250,000, but the first loan I took, I was still in college. I was minimum wage.

主持人：是的。Host: Yeah.

嘉宾：没错。我说：“给我，给我点钱，我会还给你的。”他说：“利息是10%。”我经营那家公司经营了17年，现在它是一家市值不到500万美元的公司。你在那里待了17年。Guest: Right. And I said, "Give me give me some money and I'll pay it back to you." He said, "10% interest." I ran that business for 17 years and now it's like a sub $5 million company. you stayed there for 17 years.

主持人：从某种程度上说，这是我职业生涯中犯过的最大错误。嗯，不是因为那家公司不好，而是因为它的商业模式与我想在生活中做的事情不一致。有时候你知道，我们的 CEO 查克会跟我说：“格雷格，一旦你真的上市了，就没法对商业模式做太大调整。你的商业模式就是你的商业模式。要改变商业模式，需要付出很大努力。”但当你身处服务行业时——那是一家市场研究公司，也是一家服务型企业——而我热爱软件，也热爱硅谷。我以前常来这里，还曾在俄罗斯山住过一阵子。我来到这里，就会对硅谷的人们着迷。我会想：“我想和这些人一起工作，因为他们做出的决策，以及 Windows 里的那个开始按钮，都会影响数十亿人，而我想成为做出这些决策的人。”你知道吧。Host: It was in some ways my biggest mistake in my career um because not because the company wasn't good is because the business model did not align with what I wanted to do in life and what and sometimes you know like our CEO Chuck tells me he's like grew once you actually go public um you can't alter your business model that much. Your business model is your business model. takes it takes a lot of effort to alter your business model. But when you're when you're in a services business, this was kind of a market research company and a services business. And I loved software and I loved Silicon Valley. And I used to come here and I had an apartment in Russian Hill for a while. And I would come here and I'd just be enamored with the people in Silicon Valley. And I'm like, I want to go work with these folks because the decision they make and that start button in Windows is going to be affecting billions of people and I want to be that person, you know.

呃，但我我从来没有】【：】【“】【Um, but I I never

选择学习，而非自负Choosing Learning Over Ego

我确实从中获得了同等程度的满足感。但我继续待在那里，是因为自负。这也是我要给创业者的另一个经验。我继续待在那里，是因为我自以为绝不会给别人打工。这是我自己的公司，我是自己的老板，我不想给别人打工。这是我犯过的最大错误。really was getting that same level of uh satisfaction from that. But I stayed on there because of ego. And this is the other lesson I'll give to founders. I stayed on there because I had this ego that I will never work for someone else. This is my own company. I'm my own boss. I don't want to work for someone else. biggest mistake I made

嘉宾：因为我的决定是由自负驱动的，而不是由学习驱动的。所以最后我决定去 EMC 工作，而在 EMC 雇用我的那个人现在仍然是我的教练。嗯，当时我们正在吃晚饭。Guest: because ego drove my decision rather than learning driving my decision. So finally I decided to go work at EMC and the guy who hired me at EMC is still my coach. Um, and we were having dinner.

主持人：EMC，一家大型公司，Host: EMC, huge company,

嘉宾：那是家大公司，有6万人，对吧？呃，我这是从大概25个人的公司，一下子跳到有6万人的公司。他聘我当 CTO。呃，我们就像这样隔着桌子吃晚饭。他叫 Rick Devonoodi。我对他说：“Rick，你得多给我点薪水，因为你知道，我有17年的经验。”他说：“Ju，你没有17年的经验。你是把一年的经验重复了17次。”Guest: huge company, 60,000 people, right? Uh, I'm going from like 25 people to 60,000 people. He's hiring me as a CTO. Um, and we're having dinner uh like this across the table. I'm like, his name is Rick Devonoodi. And I said, Rick, uh, you got to pay me more because, you know, I've had 17 years of experience. He said, Ju, you don't have 17 years of experience. You have one year of experience 17 times over.

主持人：来为我工作，第一年我就会给你17年的经验。Host: Come work for me and in the first year I'll give you 17 years of experience.

嘉宾：我从他身上学到的东西，比我自己经营企业的17年里学到的还多。那真是太神奇了。Guest: I learned more from him in the first year than I did in the 17 years running my own business. It was absolutely magical.

主持人：那段时间我很讨厌为他工作。Host: I hated working for him during that time.

嘉宾：但我们离开之后，我去了 Box。那时有一个人，她是我人生中的联合创始人。我觉得，你知道，她是我的运营主管。没有她，我不会接任何工作。我和她总会想：这时候 Rick 会怎么做？Guest: But after we left and I went to Box, there's this person that I is my co-founder in life. I think of like you know she is um she's my head of operations. I don't take a job without her. Her and I would always be like, what would Rick do at this point?

主持人：我们一直在问自己这个问题。所以，有这样一位优秀的导师真的很棒。嗯，后来我去了 Box。在我任职期间，我们把这项业务的规模从 2 亿扩大到了 8 亿。到最后，我负责所有产品。之后，我本来打算去一家 FAANG 公司负责广告业务。Host: And we kept asking ourselves that. And so that was a great mentor to have. Um then I went to Box. Box. We took that business from 200 to 800 million during my time there. I I ran all product by the end of it. Then I was going to go run advertising for a fang company.

嘉宾：完全和思科没有任何关系。Guest: Not at all anything to do with Cisco.

主持人：是的。我已经接受了这份工作，但他们准备给杰西的那种待遇并不适合她。她是我人生中的共同创始人。我的意思是，我不会丢下她独自离开，因为她一直陪着我。嗯，所以我当时想：“好吧，我得想办法弄清楚该怎么办。”嗯，纯属巧合的是，查克第二天就给我打了电话。Host: Yeah. I had accepted the job, but the kind of offer they were going to give to Jesse, who was with me, was not the right offer for her. And she was my co-founder in life. Like, I was not going to go anywhere without her because she stuck with me the whole time. Um, and so I'm like, "Okay, so I need to figure out what to do." Um, and literally happen stance, Chuck calls me the next day

嘉宾：他说：“嘿，你或许应该考虑来这里。”我告诉他：“嘿，听着，我已经接受这份工作了。”他说：“那家公司不管有没有你都会发展得很好，但如果你来思科，公司的发展轨迹就会根据你的作为而改变。”然后就变成了：“所以，他在 9 天内给了我们两份录用邀约。”所以你知道，这家公司愿意以非常快的节奏运作。我问他的问题是：“你会给我多大的自由度，让我专注于产品？”呃，他说：“当然，我们希望你对产品着迷。”于是我们就这么做了。不知不觉间，思科成了这样一个地方——我觉得，在我的职业生涯中，我成长最多的一段时期大概就是过去六年。Guest: and says, "Hey, you should probably think of coming here." And I told him, I'm like, "Hey, look, I'm already already accepted this job." And he said to me, he's like, "That company will do well with or without you, but if you come to Cisco, the trajectory will be shaped based on what you do." And then it was like, "So, he gave us two offers within 9 days." So, you knew that the company was willing to work at a very fast clock speed. The question I asked him was, I'm like, "How much freedom are you going to give me to just be very products?" Uh, and he said, "Yeah, we'd love you to be product obsessed." And so we did that and before you know it like you know Cisco um um became the place where I think I got some of my in my career the largest amount of growth I've had is probably in the past six years.

你会和某个像你当年19岁左右的人分享哪些经验？你知道，也许他们是在印度长大的，或者是在硅谷以外的地方长大，他们并不身处其中，而是从外部观察。我想……What are some lessons that you share with you know someone who is like you at 19 or so perhaps even growing up in in in in India or someplace outside of of the valley where they're not where they're they're looking in from the outside. I think the

雄心、团队合作与复利效应Ambition, Teamwork, and Compounding

我想说的第一课是，嗯，你的上限通常取决于你自己的雄心，而不是嗯其他任何因素。一般来说，问题不在于你的能力。我是个智力非常普通的人。嗯，我没有名校的漂亮学位。呃，但我比所有人都更有拼劲。所以现在，嗯，我的拼劲仍然比大多数共事的人都足，不是因为他们不努力，而是因为我就是觉得不能辜负他们。所以我会继续、继续坚持下去。没人能说我没有投入时间。我认为第一件事就是投入时间，放胆去想，树立非常宏大、非常大胆的目标，因为你的想象力——想象力往往比我们意识到的更容易限制我们。你知道，如果你的目标只有这么大，你可能永远到不了这里；但如果你的目标定在那边，就算你只走到这里，也已经很不错了。你知道，所以这是一个很重要的方面，值得记在心里。the first lesson I would say is um your upper limit is typically bound by your own ambition rather than um the um any other kind of factor. It's not generally your ability. I'm a very averely intelligent person. Um I didn't have a fancy degree from a fancy school. Uh but I was hungrier than everyone else. And so I still um am hungrier than most people I work with, not because they're not hungry, but because I just feel like I don't want to let them down. And so I'm just going to keep keep going at it. And there's no one who can tell me that I'm not putting in the time. And that I think is the first thing is put in the time, think really big and have very bold uh ambitions because if your imagination like imagination can tend to curtail us much more than we give it credit for you know if you have your goal that's this big chances you'll never get here but if you have your goal that's over here and if you get here that's pretty good you know and so that's a pretty important uh you know dimension to keep in mind.

第二点是，商业是一项团队运动。要为你所建立的长久关系感到自豪，在这些关系中，你确实让别人因为你的参与而取得了成功。我认为这会带来很大的助益。如果你从19岁起就这样思考，天哪，你拥有的优势会大得惊人。然后，创始人通常会听我讲的第三点，是复利的力量：你只需要做到今天比昨天进步1.27%，一年之内你就会进步100倍。如果你坚持10年，那就相当不错了。从更大的时间尺度来看，10年……The second one is business is a team sport. Take pride in longevity of relationships that you can build where you have actually made other people successful as a function of you and that I think will go a really long way and if you think that way from the time you're 19 boy the advantage that you have is tremendously high you know um and then the third area that I usually tell founders about is this notion of power of compounding like you just have to get 1 27% better today than yesterday and within a year you're 100x better. And if you do that for 10 years, like that's pretty good. And 10 years in the grand scheme of things,

主持人：对19岁的人来说，这感觉像是很长一段时间。Host: it feels like a long time when you're 19.

嘉宾：就像萨姆·奥尔特曼曾经写过的一篇博客里说的，日子漫长，但几十年很短。Guest: Like Sam Alman had this blog he had written once, the days are long, but the decades are short.

主持人：我每年每年都会读到这个。Host: I read that every every year.

客人：这是一篇精彩的博客文章。Guest: It's an amazing blog post.

主持人：这是他写过的最好的博客文章之一。没错，杰夫·贝索斯会说，对于创始人来说，能够以 20 年为时间跨度思考，而不是只看 3 个月，会实实在在地改变你做出的决策质量。你想想那些我们都敬重的真正伟大的企业家，比如黄仁勋，比如埃隆·马斯克，他们都没人指责过他们目光短浅，只看眼前。我是说，他们长期坚持、埋头苦干。唯一的区别在于，要在商业模式确实行得通的时候坚持苦干；但别像我那样，在一个并不真正契合你抱负的生意上埋头苦干。因为我经营的那家公司，很适合做一家生活方式企业。Host: It is one of the best blog posts he's written. Yeah, like Jeff Bezos will say like the ability for someone for a founder to think in a 20-year window versus a 3-month window changes the quality of decisions you make materially. And if you think about the the the really big entrepreneurs that we all respect, think about Jensen, think about Elon Musk, um none of them were accused of thinking in very short-term windows. I mean, they grinded it out for a long time. The only difference is grind it out when the business model actually makes sense, but don't do it like what I did and grind it out in a business which doesn't actually fit your ambitions. Because my business, the one that I ran, was great for a lifestyle business.

嘉宾：想要改变世界，这条路走得太糟糕了。我曾经想改变世界。你知道，想要改变世界的话，做这种生意并不合适。但如果你想把它做成一门能带来美好生活、能为家人创造财富的生意，那就太棒了。继续做下去，很好。顺便说一句，这些选择都没有错。只是我的抱负恰好和这种商业模式不相符。还有一点你必须牢记：你的抱负是否与你当下实际在做的事情相符？等你80岁回首往事时，你会不会觉得自己没有虚度光阴，会说：“你知道吗，我很庆幸自己把时间花在了这些事情上。”我觉得这是我们应该不断问自己的问题。对于年轻人，我总是告诉他们，时间过得飞快。我从没想过自己会55岁，你知道吧；也没想过别人会叫我叔叔。Guest: It was terrible to go change the world. I wanted to go change the world. You know, that's the wrong business to be in to go change the world. But if you want it to be a good lifestyle business and have it be a thing that provided wealth for your family, fantastic. Keep doing that. That's great. And by the way, none of those are wrong answers. It just happens to be that the one that I my ambition did not align with the with the model of the business which is the other thing you have to keep in mind is does your ambition align with what you're actually doing right now where you are going to get fulfilled when you look back when you're 80 and say you know what I feel very good that this is how I spent my time I think that's a question we should keep asking ourselves and for the younger folks I always tell them time's going to pass by so fast I never thought I'd 55, you know, I never thought I'd be called uncle.

我从没想过自己会有孩子。I never thought I would have a kid

主持人：我从没想过她会说：“爸爸，你老了，你知道吗，而且秃顶了。”呃，所以有时候，这些事会让你谦逊很多，提醒你一定不要把时间视为理所当然，而要把时间花在真正重要的地方。Host: and I never thought that she would say, "Dad, you're old, you know, and bald." Uh, and so like those things sometimes give you a lot of humility that you have to make sure that you don't take time for granted and spend the time in the places that actually really matter.

不要浪费你的机会Don’t Squander Your Opportunity

嘉宾：对于创业者，或者对于那些加入思科这样的组织、想像你一样推动变革的人，你有什么建议？我给你讲个故事。我……自从 1991 年离开印度后，就再也没回去过。我爸爸 2004 年去世。在他去世前，我曾回去看过他一次，只待了 48 小时。但直到 2017 年我才再次回去。所以从实际情况来说，1991 年到 2017 年间，我在印度待的时间并不多。我带女儿去了那里。我们去了泰姬陵，有个导游叫拉吉，带我们参观泰姬陵。这人看起来特别有产品思维。他对自己的产品——也就是他卖给我们的泰姬陵游览体验——了解得事无巨细。但后来不断有人从旁边经过，他就会突然用不同的语言跟人搭话：普通话、德语、法语、西班牙语。Guest: What advice would you have for entrepreneurs or for people who are coming into an organization like Cisco and trying to transform it the way that you do? I'll give you the story. So when I I I hadn't gone back to India from the time I um left India 1991 I left. My dad died in 2004. I'd gone to see him for 48 hours once before he died. But then I hadn't gone back until 2017. So for any kind of practical purpose from 1991 to 2017 I hadn't spent that much time in India. I took my daughter there. We went to the Taj Mahal and there's this tour guide. His name is Raj. He's showing us the Taj Mahal. guy seemed like a really good product guy. He knew every single detail about his product, which was the experience he was selling about the Taj Mahal. But then people would keep walking by and he would actually bust out in different languages, Mandarin and and German and French and Spanish.

然后有那么一刻，我就问他：“哥们儿，你会说多少种语言？”他说了个特别夸张的数字，差不多十二种、十四种什么的。于是我说：“这也太不可思议了。”他说：“嗯，你知道，我想尊重来到这里的人，因为他们来到这里，我不该自以为是地觉得他们会说我的语言；我应该说他们的语言。”我是说，这真的很高尚。显然是个聪明人。我心想：“这人比任何人都聪明。我当时在 Box 工作。我认识的所有同事里，他可能每天只挣十美元，而我却坐在硅谷。”为什么会这样？因为我恰好拥有一个平台，而他没有，对吧？我有教育这个平台，也有美国这个平台。And at some point I'm like, "Dude, how many languages do you speak?" He gave me like some ridiculously large number like 12 or 14 languages or something like that. And so I said, "That's that's crazy." It's like, well, you know, I want to honor the people that come here because um they come here and I should not be presumptuous enough to think that they're going to speak my language, I should speak their language. I mean, that's very noble. Clearly a smart dude. Uh I'm thinking to myself, I'm like, "This guy is smarter than anyone. I used to work at Box at the time. Anyone that I work with, he's making probably $10 a day. I'm sitting here in Silicon Valley." Uh and why is that? Because I happen to have a platform and he doesn't, right? I have a platform of education. I have a platform of America.

我拥有一个科技平台。我拥有一个平台，它给了我机会，而他却没有。因此，无论他多聪明，无论他多努力，他都很难从人群中脱颖而出。如果你有幸获得了一个平台，我唯一想对大家说的就是：不要浪费它。作为一名创始人，嗯，你知道，我们都可以把所有事情看得悲观，只看到自己缺少什么。但天哪，我们正生活在一个极其特别的时代。如果你现在正值19或20岁，或者刚刚开始职业生涯，恰逢人工智能浪潮，你知道有多少五十多岁的人愿意花多少钱，只为回到19岁吗？这是一个在科技领域活着的神奇时代，也是一个真正能对世界产生影响的好时机。I have a platform of tech. I have a platform that just gives me the affordance of opportunity that he didn't get. And so, no matter how smart he was, no matter how hard he's working, he's having a hard time breaking away from the pack. If you happen to be afforded a platform, my only ask of people is don't squander it. And as a founder um you know like we could all look at all the things as glass half empty and these are the things that we don't have but man we are all living in one of the most special times and if you're 19 or 20 or in the earlier part of your career in this moment in time during the movement in AI do you know how many people in their 50s how much money they would pay to get back to being 19 if they could? This is a magical time to be alive in tech. This is a great time to actually make an impact in the world.

如果你把时间花在与 YC 相关的事情上——我认为 YC 是最出色的组织之一，这也是我喜欢和你们共事的原因——那你就该拼尽全力，竭尽所能。不是为了别的，也不是为了赚钱，而是因为等你年纪大了，回想起自己对整个社会产生的影响时，会感到无比欣慰，那一切都值得去做。如果你碰巧把公司卖给了一家大公司，加入大公司后别变得愤世嫉俗。去学习大公司能教给你的东西，因为所有大公司曾经都在某个时刻是初创公司。如果你成功了，最终也会成为一家大公司。别贬低大公司。要真正学会如何让大公司也能带着一种……你知道的，那种闯劲去运作。And if you are spending your time associated with YC, which is one of the finest organizations I think, which is why I love working with you folks, you would be nuts not to just bust your tail and do everything you can. Not for any other reason, not for making the money, but because the impact you will have on society at large will be so gratifying when you're older that it'll be something that's worth doing. And if you happen to sell your company to a large company, don't get cynical going to a large company. Go learn things that the large company teaches you because all those large companies at some point in time were startups. And if you are successful, you will eventually be a large company. And don't dislarge companies. Actually learn how to make large companies operate with a level of, you know, kind of hutah.

这样你也确实能从中受益。带着极强的好奇心去做，无论你选择哪条路，都要知道你拥有这样一个平台。放眼全球80亿人，能拥有我们这样平台的人少之又少。And then you'll you'll actually be able to benefit from that as well. and go in with an extreme level of curiosity and regardless of which way you go but but know that you're afforded the platform and very few people in the world look at the 8 billion people very few people are afforded a platform like we have

主持人：GI，这是一个如此鼓舞人心的结尾，非常感谢你参加我们的节目。Host: GI that's such an inspiring note to end on thank you so much for joining us

嘉宾：谢谢你邀请我，真是……Guest: thank you for having me it was