Cerebras 630亿美元IPO背后的故事——创始人兼首席执行官安德鲁·费尔德曼

Netflix以前是寄DVD和信封的。

Netflix used to deliver DVDs and envelopes.

等到互联网变快了，他们就转型成了影视制作公司。

And when the Internet got fast, they became a movie studio.

对吧？

Right?

这开辟了一个完全不同的新行业，本质上是全新的东西。

It opened up an entirely new business, something fundamentally different.

速度一快，就会发生这样的变化。

That's what happens with speed.

我觉得快速发展的AI也是同样的道理。

And I think that's what fast AI does.

现在我们正在取代那些人人都看得见的东西，比如编程、设计、SaaS工具。

Right now, we're replacing things that everybody can see, like coding, design, the SaaS tools.

但一旦我们开始围绕这个进行根本性的重组，你就会看到全新的商业模式和生产力的飞跃。

But once we start sort of fundamentally reorganizing around this, you're gonna see this sort of new business models and fundamental jumps in productivity.

我特别期待那一天。

I'm eager for that.

这太棒了。

That's so cool.

今天在NoPriars，我们邀请到了Cerebras的联合创始人兼首席执行官安德鲁·费尔德曼。

Today at NoPriars, we have Andrew Feldman, the cofounder and CEO of Cerebras.

Cerebras成立于2010年代中期，最初专注于AI领域的新工作负载，特别是机器学习方向，后来迅速转向了我们今天所处的大型基础模型的高效推理领域。

Cerebras was founded in the mid two thousand tens to focus on new workloads for AI, particularly the machine learning world, and then has made the transition in a very fast inference for the foundation model world that we live in today.

Cerebras最近刚上市，目前在股市上的估值约为630亿美元。

Cerebras recently went public and is currently worth about $63,000,000,000 in the stock market.

所以，安德鲁，感谢你做客NoPriers。

So, Andrew, thank you for joining us on NoPriers.

这真是我的荣幸。

Oh, what a pleasure.

很高兴再次见到你们。

It's good to see you guys again.

是的

Yeah.

首先，恭喜你。

So first of all, congratulations.

你们公司Cerebras刚刚上市了。

So your company Cerebras just went public.

截至今天，市值达到了六千亿美元，非常了不起。

As of today, it's a $60,000,000,000 market cap, which is pretty amazing.

真的很了不起。

Pretty amazing.

是的

Yeah.

我记得你大概一两年前来过我们的节目，是早期的某一期，那次和你聊天非常愉快。

And I think you were with us a year or two ago on the show in one of the earlier episodes, it was a pleasure to talk to you then.

今天能再次邀请你来，我们非常高兴。

Obviously, we're very excited to have you on today.

你能跟我们聊聊，从那以后公司是怎么发展的吗？也顺便给我们的听众做个提醒，你们到底在做什么，专注在哪些方面？

Can you tell us a bit how the business evolved since that time and what you folks just a reminder for our audience what you do, what you're focused on,

你们接下来打算怎么推进？

how you're moving forward?

我们打造的是人工智能计算机。

We build AI computers.

没错。

Right?

专门设计并优化用来加速人工智能任务的计算机。

Computers designed to and optimized to accelerate AI workloads.

目前我们在推理速度上遥遥领先，不是快一点，而是快很多。

And right now, we're the fastest at inference, not by a little bit, but by a lot.

比显卡快十五到二十倍。

Fifteen, eighteen, 20x faster than GPUs.

大概从2025年开始，人工智能模型变得足够智能，真正派上用场了。

And so what happened was starting at about 2025, AI models got smart enough to be useful.

人们开始使用这些系统，我们做AI训练，也用它来做推理。

People began using them, and you know, we make AI with training, and we use it with inference.

随着大家越来越多地使用，它逐渐融入了日常的工作流程。

So as people began to use it, it began to sort of be integrated into their day to day work.

速度变得至关重要，我们的订单多到根本忙不过来。

Speed became fundamentally important, and we were just crushed with demand.

这种速度优势是全面的，还是只针对某些特定场景？

Is this faster across the board, is this specific use cases?

是全面的更快。

Faster across the board.

不管是大型模型还是小型模型，美国的模型还是中国的模型，万亿参数的，还是十亿参数的，全都快得多。

Big models, small models, US models, Chinese models, trillion parameter models, 1,000,000,000 parameter models across the board.

到了年底，我们和OpenAI签了一笔交易，是硅谷历史上最大的交易之一，金额超过两百亿美元。到了三月，我们又和AWS达成协议，未来他们的数据中心将部署我们的系统。

And then what happened was at the end of the year, we signed a deal with OpenAI, of one of the biggest deals ever in Silicon Valley, sort of north of $20,000,000,000 And then in March, we signed an agreement with AWS where we will be deployed in their data centers going forward.

那一年半简直像坐过山车，我们一直在拼命追赶产能，想方设法满足需求。

And so it was just a whirlwind year and a half of chasing the chasing supply and trying to trying to sort of meet

需求。

the demand.

过去一年，尤其是最近一年半，到底发生了什么变化？

And what what shifted in the year and in the last year and a half?

是制造产能提升了？

Was it the ramp in manufacturing?

是新芯片设计的突破？

Was it a new chip design?

还是别的原因？

Was it something else?

你能给大家解释一下吗？

Could you help educate What folks on

其实我们造出了一台特别特别快的机器，但很长一段时间都没人在意。

happened was we built a really, really fast machine, and for a long time nobody cared.

对吧？

Right?

那是因为，呃，不好意思打断一下

That's because Actually, forgive

恕我直言，很多人反对，说这只是一个奇怪的架构。

me for saying so, but a lot of people objected and said this is just a weird architecture.

他们觉得这走错了路。

They called it wrong.

连Cerebras都说是错的。

Like Cerebras called it wrong.

是的。

Yeah.

没错。

Yeah.

确实如此。

They did.

我觉得，要想有根本性的突破，就不能只在原有架构上做改良。

I think to be radically better, right, you can't build something that is a similar architecture.

对吧？

Right?

想比GPU提升十五到二十倍，光靠对现有架构做点小改动是做不到的。

You're not going to get 15 or 20 times better than the GPU with a minor modification to their architecture.

这一点在各行各业都差不多：如果你真想实现根本性的突破，设计就必须不一样。

And that's probably true across the board, that if you're going to aspire to a radical improvement, your design has to be different.

从一开始，我们就选择了晶圆级方案，也就是说，我们造出了一个四万六千平方毫米的芯片，大小跟餐盘差不多，而其他人做的芯片都只有邮票那么大。

And from the beginning, you know, we chose wafer scale, which means we build a 46,000 square millimeter chip, a chip the size of a dinner plate, whereas everybody else is building chips the size of postage stamps.

他们说我们疯了。

They told us we were out of our mind.

这根本行不通。

It would never work.

他们还列出了一大堆理由，说这根本不可能实现。

They listed reasons why it was impossible.

但到了2019年，我们证明了这是可以做到的。

But in 2019, we we proved it was possible.

我们开始交付产品，然后不断优化，再继续优化。

We began delivering it, and we improved on it, and we improved on it.

但当时人工智能还只是个新鲜玩意儿，我们跑得快也没用。

But we were fast when AI was a novelty.

当一个东西还很新奇的时候，没人会在意你快不快，因为根本没人用。

And when it's a novelty, nobody cares that you're fast because it's not being used.

从2023年到2025年初，大家嘴上都谈AI，但没人真把它用在日常工作中。

And so from about 2023 to the beginning of '25, sort of people pointed at AI, but nobody used it every day in their work.

对。

Mhmm.

一旦你每天都要靠它干活，速度就绝对不能拖后腿。

And once you use something every day in your work, it can't be slow.

你说，你愿意等一个网页加载多久？

I mean, how how long will you guys wait for a website to resolve?

你根本留不住用户。

You'll have no attention.

没错。

Right.

说得太对了。

That's exactly right.

确实就是这样。

That that's exactly the way it is.

你说，慢速搜索的市场能有多大？

I mean, how big is the market for slow search?

零。

It's zero.

拨号上网的市场有多大？

How big is the market for dial up Internet?

也是零。

It's zero.

慢速推理的市场也会是这个规模，但我们得等到它足够智能、真正有用才行。

That's how big the market for slow inference will be, but we had to wait until it was smart enough to be useful.

这件事发生在2025年。

And that happened in 2025.

所以才出现了像Cognition、Cursor、Lovable这些公司的需求爆发，它们全都开始疯狂增长。

And that's why you got this sort of explosion of demand in companies like Cognition and Cursor and Lovable and just all these others that began ramping extraordinary.

你们投资的很多公司，包括OpenAI在内，都在迅猛发展。

Many of the ones you guys have invested in are ramping like crazy, OpenAI and others.

而我们正好推出了对的产品。

And we were right there with the right product.

我记得大概是2016年左右第一次见到你。

I think I first met you back in 2016 or something like that.

那时候，人们还不觉得AI这个词听起来怪怪的，对吧？

And at the time, people weren't like, saying AI sounded weird, right?

你当时谈的是机器学习，模型还都是卷积神经网络、RNN，还有刚冒头的GAN之类的东西。

You were talking about machine learning, and the models at the time were convolutional neural networks and RNNs and just the emergence of GANs and things like that.

我们那时候还在努力区分椅子和猫。

We were trying to tell the difference between a chair and a cat.

对吧？

Right?

那是Quocly那会儿对人脸的大规模识别。

That was Quocly's great swarm of his face.

比如区分猫和椅子。

Like a cat and or a chair.

天啊。

Woah.

看看我们走到了这一步。

Look how far we've come.

简直难以置信。

I mean, it's unbelievable.

是的。

Yeah.

对。

Yeah.

你觉得是什么让你有远见，敢逆着市场去做这件事？

What do you think gave you the foresight to build against the market?

你说得对，很多人都觉得这个市场会很重要，但你比其他人更坚定，毕竟你真正在这个领域创办了公司。

Because to your point, think a lot of us believed in this market would be really important, and you more than others, right, since you actually started a company in it.

但市场真正壮大起来花了挺长时间，就像你现在说的，才变成这么大的应用场景，大家开始真正关注推理速度这些事。

But then it took some time for the market to really expand to the point where, to your point now, it's this massive use case, people really care about speed of inference and other things.

那时候，是什么让你有这么强的信心去干这件事？

What gave you the conviction back then to do this?

是远见、合适的联合创始人，再加上一点傲气和一点运气。

Combination of vision, the right cofounders, and a little bit of arrogance, a little bit of luck.

我们当时就看到AI正成为一种新的计算负载。

You know, we saw AI on the horizon as a new workload.

作为计算机架构师，新的负载就意味着机会。

And as computer architects, new workloads are opportunity.

对吧？

Right?

进入x86领域太难了，那里根本没什么新东西，几十年都没变过。

It's very, very hard to to to enter in the x 86 world, right, where there's not nothing new is happening there, and nothing has happened for generations.

但你看，图形计算兴起的时候，出现了独立显卡，有了NVIDIA；移动计算爆发时，ARM就崛起了。

But, you know, when graphics emerged, you got the discrete GPU, and you got Nvidia, and when the mobile compute hit, you got ARM.

有意思的是，英特尔、AMD，还有那些本该最有优势的大公司，全都没能分到一杯羹。

And it was interesting that not Intel, not AMD, not all sorts of people who you would have thought have been really well positioned to win in that business, they all got no share.

所以我们知道，这种新负载会消耗海量的计算资源。

And so we knew that that that this new workload would eat a lot of compute.

它需要一套全新的架构，专门设计的架构，而且必须跟现有的完全不同。

It would require a new architecture, dedicated architecture, and that ought to be very different.

这个架构不能是现有技术的衍生品。

The architecture could not be a derivative of what's existing.

这些就是我们下注的核心判断，完全跟主流反着来。

Those were our big bets, and they were a 100% contrarian.

结果证明，我们全对了。

And they turned out to be dead right.

在过程中，你有没有哪一刻真的怀疑过这事能不能成，考虑到……

Were there moments where you just doubted whether this would work, given the

确实不一样。

different for sure.

是的。

Yeah.

有一段时间，我们在解决一个前所未有的问题。

We had a period you know, we're solving a problem that had never been solved before.

换句话说，在计算机行业七十多年的历史里，一直有人尝试制造晶圆级的产品。

I mean, there'd been efforts across the entire seventy year history of the compute industry to build a wafer scale product.

事实上，基因·阿姆达尔——我们这个领域的奠基人之一，计算界的传奇人物之一——就彻底失败了。

In fact, Gene Amdahl, sort of one of fathers of our field, one of the guys on Mount Rushmore of compute, failed miserably to do it.

从2017年年中到2019年年中，有一段时期我们根本做不出来。

We had a period between about 2017, middle of twenty seventeen and middle of twenty nineteen, where we couldn't build it.

那时候我们每个月要花大约八百万美元。

We were spending about 8,000,000 a month.

你每六周开一次董事会，都说：这东西我造不出来。

You're having board meetings every six weeks saying, I can't build it.

不行，还是不行。

No, it's still not working.

没错，‘唉’说得太对了。

And right, Oof is right.

我的意思是，投资者愿意投这么多钱，还这么坚信，这真不容易。

I mean, that's a huge amount of money and a huge amount of conviction your investors have.

每次做失败分析，我们都能从中学会一点东西。

And each time we did a failure analysis, we got a little bit better at it.

我们每次都能进步一点点。

We got a little bit better at it.

到了2019年夏天，我们终于成功了，它开始运转了。

And then in the summer of nineteen, we yielded it, and it began to work.

第一次成功时，我们坐在洛杉矶洛斯阿尔托斯市中心一个临时改造成的办公室里，那地方根本不是给搞硬件的人设计的，我们盯着一台电脑，那场景简直无聊得像看油漆干——对。

And the first time, we were sitting in a little makeshift shift office in Downtown Los Altos in a building that was not designed for hardware guys, and we're staring at a computer, which is about as exciting as watching paint dry Mhmm.

它终于能用了。

And it's working.

我们俩整整半小时都说不出话来。

And we just we couldn't speak for half an hour.

对吧？

Right?

从来没人做成过这事，可我们做到了。

It's like, nobody had been able to do this, and it's working.

这是我们做出来的。

And we did this.

太了不起了。

That's amazing.

这说的是技术层面，还有市场层面呢。

Because that's the technical side of it, then there's a market side.

对吧？

Right?

在市场方面，正如你所说，要让这些工作负载变得真正重要，也花了不少时间。

And also on the market side, to your point, it took time to get to the point where these workloads were really important.

那有没有哪个时刻，你怀疑过这个市场根本不存在？

So were there moments where you doubted whether the market existed?

我们把问题解决了，还解决了计算机行业最难的问题，可没人在意。

You know, we we solved it, and we solved this sort of the hardest problem in the computer industry, and nobody cared.

真的一个人都没有。

Nobody.

第一代产品，我们可能就卖了十几台。

It was like, you know, the first gen, we might have sold a dozen.

第二代大概卖了三百台，到了第三代，现在销量已经达到了数万台。

The second gen, we probably sold 300, and now we're selling in a sense of tens of thousands in the third gen.

我们有两三年时间是走在市场前面的，哪怕我们的速度快得惊人，也没人关心。

We had a two or three year period where we were ahead of the market, and absolutely nobody cared that we were blisteringly fast.

你们找到了一些非常特别的早期客户，对吧？

And you found some pioneering customers that were, like, atypical in terms of a starting point, right?

有一些国家机构确实提前采购了。

Were some sovereigns who really bought ahead.

你是怎么应对这段需求滞后时期的呢？

Like, how did you think about being resilient to this period of being ahead of demand?

我觉得新的计算机架构已经铺好了一条路。

I think there's a path that has been laid down by new computer architectures.

通常我们会先从超级计算领域切入，因为那里的用户只看重速度，根本不介意软件是不是还不成熟。

And often you begin in the supercompute world because those guys love speed, they don't care if your software is immature.

所以我们在那里一路横扫。

And so we sort of ran the table there.

我们在阿贡国家实验室、劳伦斯利弗莫尔实验室、桑迪亚实验室，还有欧洲的LRZ并行计算中心都赢得了订单。

We won at Argonne National Labs and at Lawrence Livermore and at Sandia and in Europe, the European Parallel Computing Center at LRZ.

之后我们又拿下了石油天然气行业和制药行业的几家客户，他们历来都极度依赖高性能计算。

So we ran the table there, and then we won some guys in the oil and gas space, and we won some guys in pharma, all of whom have long histories of using extraordinary amounts of compute.

但过去一直存在一个巨大的鸿沟，因为这些客户都无法提供足够的量来推动产品进入主流市场。

But then historically, there's this giant chasm because none of them provide the volume to get to mainstream.

我们还赢得了一家主权机构——G42，他们成了我们的战略伙伴，也是亲密的朋友，还给我们下了十亿美元的订单。

And we won a sovereign, a G42, and they became strategic partner and close friends, and they placed a billion dollar order on us.

有了这笔订单，我们才真正实现了公司的转型。

And with that, we were able to sort of transform the company.

我们得以重新调整供应链。

We're able to change our supply chain.

我们能够部署足够大规模的集群，从而在真实环境中进行全面测试。

We're able to deploy equipment in big enough clusters that that we could battle test at scale.

你知道，硬件领域的一个难题是，你的质量测试实验室不可能和你想要服务的客户规模一样大。

You know, one of the challenges in hardware is your QA lab can't be as big as some of the customers you wanna deploy to.

对。

Mhmm.

对吧？

Right?

你总不能把价值上亿美元的设备全塞进自己的测试实验室吧。

But you can't put a $100,000,000 in your QA lab worth of your own gear.

他们和我们紧密合作，开始让我们为他们训练模型。

And they worked with us, and we began training models for them.

我们也开始为他们做推理服务。

We began doing inference for them.

他们真是个难得的合作伙伴。

They've been an extraordinary partner.

这位是G42的首席执行官彭先生，还有他的董事长石德南。

This is Peng, who's CEO G42, and his chairman, Shi Tak Nun.

我们再也找不到比他们更好的合作伙伴了。

We couldn't ask for better partners.

所以当OpenAI和AWS出现的时候，我们已经具备了相应的产能。

And so we we were able to when OpenAI came along, when AWS came along, we we had the capacity.

我们早就准备好了。

We were ready.

对吧？

Right?

我们已经经过实战检验了。

We'd battle tested.

我们差不多跨过了那道鸿沟。

We'd sort of gotten over the the chasm.

我们早就搭好了桥梁，所以能应对需求。

We we'd had a bridge, and so we could meet the demand.

没错。

Yeah.

我觉得这种路径依赖在这个领域常常被低估了，毕竟从几亿美元的订单，一下子跳到两千亿的待办 backlog，中间肯定得有人一步步把路走通。

I think that kind of path dependence is sometimes undervalued in this field because the ability for you to go from a, you know, like tens, $100,000,000 order to 20,000,000,000 of backlog, Like, there's gotta be there's gotta be something in the middle as somebody work.

对。

Yeah.

这是多年积累的成果。

It's years of work.

而且说实话，我觉得很多听众可能都做软件的，你们确实能快速扩张。

And and, you know, it's I I think often, and I'm sure many of your listeners are in the software world, and and you guys can scale so fast.

对的。

Mhmm.

对吧？

Right?

但当你在打造产品的时候，你得跟制造伙伴确认细节，比如尺寸啊、工艺啊这些。

But but when you're building things, right, you you have to you wanna double, you gotta call your manufacturing partner, your Centimeters.

他们得解决供电问题。

You gotta they have to find power.

他们得把厂房搞定。

They have to win the building.

他们得增加生产线。

They have to add more lines.

他们得制作测试工装。

They have to make test fixtures.

对吧？

Right?

每一个步骤都需要实实在在的时间和精力去推进。

Each step takes real time and effort to grow.

今年我们打算把制造能力提升十倍。

We're going to try to increase manufacturing 10x this year.

对吧？

Right?

这大概是硬件行业历史上最快的增长速度了。

That's about as fast as anybody in the history of hardware.

而且你们的软件栈也占了绝大部分。

It's also majority of the software stack for you guys.

这可是更大的规模。

That's more scale.

对吧？

Right?

还记得我们刚创业的时候，萨拉，我的联合创始人之一，我印象很深。

You know, when when we started the company, Sarah, our one of my cofounders I do remember.

我知道。

I know.

我们之前已经向你们展示过了。

We we presented to you.

我的一位联合创始人说，安德鲁，做个编译器得花十年。

One of my cofounders said, Andrew, it's gonna take about ten years to build a compiler.

我说，不可能。

I said, no.

这太荒唐了。

That's crazy.

这完全是大公司才说的话。

That's big company talk.

我们五年就能搞定。

We can do it in five.

得花十年左右。

It takes about ten years.

对吧？

Right?

结果是这样。

It turns out.

开发编译器确实花很长时间。

Takes a long time to build a compiler.

这是一套极其复杂的软件系统。

It is an extraordinarily difficult piece of software.

现在我们已经搭建起一套不错的软件栈。

And now we've got a good software stack.

我顺便问一下，既然十多年来你一直坚信这场变革会发生，

Can I ask you as an aside, actually, just because you have, for more than a decade, believed that this revolution is going to happen?

那么AI生成代码对Cerebras公司内部的实际帮助有多大？

How much is all of this AI generated coding relevant for Cerebras internally?

非常大。

Hugely.

八个月前，我们在工程上花在令牌上的钱还没到一千块，现在大概已经涨到两万五到三万了，增长得特别快。

I would say that that, you know, eight months ago, we weren't spending a thousand dollars in engineering on tokens, and we're we're probably at 25 or 30,000 right now, and it's ripping.

我觉得这并不是对所有人都有用。

I I think it's not useful for everybody.

我觉得这确实是实情。

I I think that's the truth.

确实有一部分人，他们的思维方式特别适合这种模式。

I I think there are some some people who have sort of the perfect mindset for it.

对吧？

Right?

他们同时运行着八到十个代理，全天候不间断。

And they are running eight or 10 agents, seven by 24.

他们的编程方式已经转变成以管理代理为核心，连怎么做质量测试都交给代理来想。

They've moved their coding style to being one in which they govern agents, whether they think about how to QA.

所以他们专门设了一个质量测试代理在跑。

So they've got a QA agent running.

他们会思考如何弥补代码模型的一些不足。

They think about how to sort of remedy some of the weaknesses in the coding models.

对吧？

Right?

他们写的代码往往很长。

They're often verbose.

经常把注释都删掉了。

They often cut out comments.

他们真的深入思考过，这种问题正好契合他们的思维方式。

They've really thought about, and it's a type of puzzle that the perfect fit for their mind.

他们已经从原来的一流高手，变成了百倍效率的高手。

And they've gone from being sort of 10x guys to being 100x guys.

我觉得我们其他人，包括我自己，还在艰难地摸索。

I think the rest of us, myself included, we're sort of limping along.

我们正试着弄清楚，怎么把这套方法用在各自的工作里——不管是当CEO、CFO、会计，还是做市场。

We're trying to figure out how we can make it work for our different jobs, for being the CEO, for being the CFO, for being accountants, for being in marketing.

但对少数人来说，这真是个利器。

But for a small number, it is such a tool.

至于其他人，我们会让他们看看别人是怎么做的，什么是最佳实践。

And then the rest, we try and show them what others are doing, what best practices are.

你们现在大概有八百人了？

You're about 800 people now?

八百到八百五十人左右。

800, eight fifty, yeah.

人均市值真高啊。

It's a lot of market cap per person.

我喜欢这句话。

I like that.

没错。

Yeah.

太棒了。

That's great.

总的来说，这是一个不错的指标。

It's a good metric overall.

当你思考接下来该往哪里走，比如怎么把业务做大、制定战略方向时，你预测的未来是什么？除了送货之外，我们还能往哪儿发展？

When you think about where to go from here, you know, making business bigger, strategic directions, like what you predict, and where can you go from think we Besides delivery.

当你手头的积压订单超过两千亿美元时，每天的交付能力就变得至关重要。

Well, when you've got a backlog that's north of $20,000,000,000 delivery is pretty important every day.

我觉得我们还得继续保持无所畏惧的态度。

I think we have to continue to sort of be fearless.

很多公司发展到一千人、两千人甚至三千人的时候，就会停止像以前那样大胆冒险，对吧？

I think one of the malaise of companies as they get to 1,000 to 2,000, 3,000 people is they stop taking the type of risks that they were taking before, right?

你从一个无所畏惧的工程文化，慢慢变成只想着‘下个版本能交付什么’。

You move from being a fearless engineering culture to sort of being, what can we get in in the time frame of the next rev?

我觉得这种转变非常有害。

And I think that's extraordinarily damaging.

我们非常自豪能坚持做这种无所畏惧的工作。

And we take such pride in doing fearless work.

我们想招那些敢做敢为的人。

We want to hire people who do fearless work.

我们要守护这种文化：宁愿在追求非凡时失败，也不愿在平庸中成功。

We want to kind of sort of guard that culture that says we would much rather fail in pursuit of the extraordinary than succeed in the ordinary.

这种做法太糟糕了。

That is a horrible thing to do.

所以这些才是让我担心的事。

And so those are some of the things that worry me.

我觉得招聘确实是个关键问题。

I think recruiting, right?

你有那么多职位空缺，很容易就妥协，很容易为了填满位置而将就。

You have so many openings, and it's so easy to settle, and it's so easy to just try and put a butt in the seat.

是的，还不错。

Yeah, pretty good.

赶紧把人招进来吧。

Let's get that butt in the seat.

这简直就是自取灭亡。

I mean, that is death.

所以我们非常慎重，我每天都会花不少时间亲自跟候选人交谈。

And so we think really hard, and I spend a meaningful part of every day in talking to candidates.

这些事是我每天都会担心和反复思考的问题。

Those are things that sort of I worry about and I think about every day.

我们有很多创始人和领导者，他们听这个播客，正在思考：自己的生意虽然成功了，但眼下要等市场回暖，或者怀疑自己是否还走在正确的路上。

We have a lot of founders and leaders who, you know, listen to the podcast who are thinking about maybe they have a successful business and they're managing through the period of waiting for the market or trying to figure out if they're still right.

他们在考虑如何从八百人扩张到几千人。

They think about how to hire from 800 to several thousand.

我们之前聊过，当你在问自己‘我是否还适合这个十年’的时候，该怎么管理自己的心态。

There we talked about the managing of your own psychology when you're like, am I right for this decade?

在这么长一段时间里，没有外界反馈，你该怎么留住并激励员工？

How did you like keep and motivate employees when there wasn't external feedback for this long period of time?

首先，我非常理解他们的处境。

Well, first, I have empathy for them.

我是说，当CEO真的特别孤单。

I mean, being CEO is an extraordinarily lonely thing.

你在打造一家公司，真正在打造一家公司。

And you're building a business, you're building a business.

你们都明白，当领导就是孤单，而且一点都不轻松。

You guys know this, that being a leader is lonely, and it's not easy.

没人愿意说这个，尤其是我们这种喜欢解决问题的人——偏偏专挑别人说解决不了的难题下手。

And people don't like to say that, especially for those of us who like to solve problems, specifically the problems everyone else says can't be solved.

那种不服气的感觉，反而成了你前进的动力。

You sort of you you gain fire from that chip on your on your shoulder.

对吧？

Right?

别人说这事办不成，你心里就憋着一股劲：我偏要把它办成。

When when they say it can't be solved, you say in your head, you can't solve it.

没错。

Right.

我觉得挺好的。

I thought it was just fine.

不是的。

No.

没错。

That's right.

完全正确。

That's exactly right.

你知道的，你是一家顶级风投公司。

You know, you a top venture firm.

你希望按自己的方式来做。

You wanted to do it your way.

对吧？

Right?

所以你就选择按自己的方式走了。

And so you stepped out doing it your way.

你对自己说，我能做成这事，但这并不容易。

And you say to yourself, I can do this, and it's not easy.

这是一方面。

And that's one thing.

另一方面，你得真正热爱这个过程。

The other thing is you have to love the journey.

对吧？

Right?

如果你不喜欢折腾的过程，这些事根本没法坚持下去。

These things we do are too hard if you don't like the building.

对吧？

Right?

你做这些只是为了赚钱，那可太糟糕了。

That you do this for the money is a horrible thing.

想赚钱的话，比这容易多了，何必非得去创造什么了不起的东西，还跟英伟达这样的巨头硬碰硬呢？

There are way easier ways to make money than than trying to create something extraordinary and compete with somebody as strong as as Nvidia.

这条路一点都不好走。

That is not the easiest path.

你得真心喜欢当那个大卫。

You gotta love being a David.

对吧？

Right?

我是个专业的大卫。

I'm a professional David.

这是我第五次创业了。

This is my fifth startup.

我一直在跟歌利亚较量。

I compete against Goliath.

这正是我谋生的方式。

That is what I do for a living.

我常想，我们每赚一块钱、每一百万、每一十亿，要是没有我们的智慧，对方光靠实力早就把一切抢走了。

And I think to myself that every dollar, every million dollars, every billion dollars we sell, if it wasn't for our brains, their muscle would have taken it in a heartbeat.

你得喜欢这一点。

And you gotta love that.

如果你不喜欢这一点，那这条路会非常漫长。

And if you don't love that, it's a very long road.

我觉得关于什么时候该放弃一件事，世界上有两种看法。

When do you think because there's sort of two views of the world in terms of when to give up on something.

一种观点是不管怎样都要坚持下去，希望事情最终能好起来。

And one argument is just going no matter what, and hopefully things work out or eventually they will.

另一种看法是，你应该不断重新评估自己走的这条路是否正确。

The other view of the world is, you should be constantly reassessing whether the journey you're on is the right one.

有时候，真正明智的选择恰恰是放手。

And there are some moments where actually giving up is the smartest possible thing you can do.

你对这个怎么看？

What's your view on that?

或者你觉得什么时候才是放弃一件事的合适时机？

Or how do you think about when's the right time to give up on something?

当你已经明确列出了获胜所需的所有假设，而每一个都被证明不成立时，那就明显是该放手的时候了。

I think it is clearly the right time to give up when you've laid out a set of hypotheses about what it's gonna take to win, and they all come back negative.

对。

Yeah.

但我看到很多人是一步步来的。

But I see people kind do this sequentially.

对吧？

Right?

他们说，我再试一次就行，结果试了还是不行。

They say, oh, I just need to test one more thing, and they test it it doesn't work.

那我就再试一次。

So I need to test one more.

于是呢

And so

这种心态就像滑坡，一发不可收拾。

the slippery slope is a beast.

是的。

Yeah.

在所有事情上，在道德抉择中，在你的生活中，滑坡效应都存在。

The slippery slope in all things, in ethical situations, in your life.

我的意思是，滑坡效应真的需要你时刻警惕。

I mean, the slippery slope is really something you have to guard against.

对吧？

Right?

我觉得有时候，有其他前CEO或者经验丰富的创业者站在你这边，提醒你：一年前你还说，只要做到这一步，你就不会这样，他们能帮你拉回来，避免你滑下去。

And I think sometimes having other former CEOs or other really seasoned entrepreneurs who are on your side and who can share with you remember a year ago, you said, if you got to this point, you didn't have this, and to remind you, so so they pull you back off that slippery slope.

对吧？

Right?

他们说，就像那句老话——青蛙在温水里，你以前不是说过，要是水热到这个程度，你就该跳出来了吗？

They they said, you know, the old frog in the the warm water thing is like, you said if it got this hot, you were gonna get out.

对。

Yeah.

然后温度一点点继续升高。

And it slowly kept getting warmer.

所以本质上，是靠别人来帮你真正履行责任，

So it's basically, can other people keep you effectively accountable to

没错。

That's right.

朝着正确的方向前进。

Directions.

对你自己的想法负责。

Accountable to your own thinking.

是的。

Yeah.

如果你明白问题出在哪里，能清楚说出哪些地方必须改，是吧？

If you understand why it's not working, right, if there are some things that that you can articulate that have to change Yeah.

只要能为这些改变设定一个时间期限，事情就能有起色。

In order for it to work, and and you can put some sort of time frame on it.