埃隆之书，作者埃里克·乔格森

当他设定截止日期时，他会说：我要选一个我有50%概率能完成的截止日期。

When he sets deadlines, he's like, I want to pick a deadline that I'm 50% likely to make.

我们会错过一半的截止日期，但我完全接受这一点，因为这意味着我们正在以最快的速度推进，而且我们还会完成一些没人相信我们能做到的截止日期。

We're going to miss half our deadlines, and I'm totally fine with that because it means we'll be moving as fast as we possibly can, and we're going to make some deadlines that nobody thought we were going to make.

如果我们100%都能完成截止日期，那说明这些日期定得太宽松了，因为我们总会耗尽给自己分配的所有时间。

If we're making 100% of our deadlines, then they're way too far out because we'll always consume at least the amount of time that we give ourselves.

我认为人们普遍存在一种偏见，尤其是当你的工作岌岌可危，或者想在同事面前表现得好，或者只是希望公司顺利运转时，你都不想被看作是失败者。

And I think people have this inherent bias, especially like if your job is on the line or if you want to like look good in front of your peers or if just want your company to work, you don't want to be seen as failing.

你做了那么多决策，工程师尤其如此——我不希望我的产品失败，不希望我的部分出问题。

Like you make so many decisions, engineers in general, like I don't want my product to fail, don't want my part to fail.

而他所倡导的这些文化以及这些格言，其核心恰恰是：你应该去失败。

And he's so much of what these cultures reinforce and these sort of maxims reinforce is like, you should be failing.

他说，他招聘人的时候会说：如果你说不出你搞砸过的四件事，那你根本没在做真正重要的工作。

He says when he hires people, like, if you don't if you can't tell me the four ways you fucked something up, then you weren't the one doing the real work.

我刚刚还在找这句话

Was literally searching for that quote

是的

Yeah.

对，这是个很好的观点。

Right is a great one.

但所有这些都指向一点：如果你不把你的组织设计成能够承受这些小失败，你就是在制造脆弱性和低效——这正是反脆弱理念的核心：如果你不能在过程中学会失败并快速迭代，就无法逼近更好的产品。

But it all of those kind of line up behind, like, if you're not engineering your organization to have these small failures, you're creating this is the antifragile idea, you're creating this fragility and this inefficiency in the global thing if you don't learn to fail along the way and iterate extremely quickly towards a better product.

我曾经和迈克尔·戴尔聊过，那是节目最早的几期之一，他也谈到了这一点。

I had a conversation with Michael Dells, one of the first episodes of the show, and he was talking about that too.

他说：当你在用新技术和新商业模式开创一项新业务时，根本没有任何现成的指南。

He's like, Listen, when you're inventing a new business, using new technology and a new business model, there's no playbook.

你不可能找到一本书，告诉你‘这是怎么做’。

You can't go read a book that says, Here's how you do it.

而且你甚至不能从相关行业挖人，因为他们只会用他们过去的方式去做。

And he's like, And you can't even hire people from adjacent industries because they're just going do it the way they used to do it.

所以他觉得，唯一的方法就是自己亲自去尝试，我觉得他用的词是‘亲自摸索’。

So he's like, only way to do it is to experiment into I think he used the word into it yourself.

他说，你有一个假设，然后去验证这个假设。

He's like, You have a hypothesis, and then you test that hypothesis.

他想表达的是，因此，你反而希望犯错。

And the point he was making is like, so therefore, you want mistakes.

你实际上希望犯错，因为这才是你学习的方式，但你要让错误尽量小，而且不要一再犯同样的错误——这再次印证了你刚才关于马斯克的说法。

You actually want to make mistakes because that's what you learn, but you want to make them small and you don't want to make the same mistake over and over again, which I, again, is like an echo of what you were just saying about Elon.

我想这与我眼前这本书中的一个观点有关，那就是马斯克的公司希望用现实作为验证工具。

I guess this ties to the part of the book that's right in front of me where it's clear that Elon's companies, he wants to use reality as a validation tool.

所以他提到了这一点。

So he talks about that.

他说，接近真相非常重要。

He's like getting to the truth is really important.

因此他说，在商业和个人生活中，一厢情愿会引发很多错误。

So he says in business and personal life, wishful thinking causes a lot of mistakes.

你必须追问某件事是否真实。

You have to ask whether something is true or not.

妄想是一种自然的人类倾向。

Wishful thinking is a natural human tendency.

这些话都是埃隆的原话。

Again, these are direct quotes from Elon.

要区分出是坚信一个新想法并坚持到底，还是在追逐一个不切实际的梦想，这很具有挑战性。

It is a challenge to tell the difference between believing in a new idea and persevering or pursuing an unrealistic dream.

任何初创企业的真正考验在于它如何应对逆境并做出调整。

The real test of any startup is how well it responds to adversity and adapts.

大多数事物在起步时都显得毫无道理。

When most things start out, they don't make much sense.

但只要你能快速适应，就能让公司运转起来。

But as long as you adapt quickly, you can make the company work.

坚韧不拔并极度专注于真相，是极其重要的。

Being tenacious and super focused on the truth is extremely important.

你应该从所有来源寻求反馈。

You should look for feedback from all sources.

以现实为师的迭代过程，也是马克斯·奥尔森文章中的另一个精彩观点。

The iteration towards reality as the teacher is it was another, like, great element of Max Olson's essay.

纳瓦尔对此也表达得非常清晰，就是将大卫·多伊奇的思想应用到企业建设中。

And Naval is super articulate about this as, like, applying David Deutsch's ideas to company buildings.

作为一家公司，你的目标是发现新知识，这意味着要与现实互动，然后将其转化为产品。

Like, your goal as a company is to discover new knowledge, which means interact with reality, and then instantiate it into a product.

这种迭代和实验的循环，就是你检验现实的方式，对吧？

And this cycle of iteration and experimentation is how you check against reality, right?

所有知识都是这样创造出来的。

This is how all knowledge is created.

就像基因突变与自然选择。

It's gene mutation and selection.

就像科学假说与实验验证。

It's scientific hypothesis and testing.

但人们通常不会从组织创造了多少新知识、进行了多少实验、失败能多快多低成本地发生这些角度来思考自己的组织。

But people don't usually think about their organization in terms of like how much new knowledge is being created, how many experiments are we running, how quickly and how cheaply can the failures be created?

是的，这里还有另一句精彩的话。

Yeah, here's another great line.

如果你的信念与火箭成功入轨相矛盾，那么火箭就无法入轨。

If you have beliefs that are incompatible with the rocket getting to orbit, the rocket will not get to orbit.

物理是一个严厉的裁判。

Physics is a harsh judge.

我会自己记下这一点。

I'll note that myself.

一遍又一遍地重复。

Repeats over and over again.

物理是一条法则。

Physics is a law.

其他的一切都只是建议。

Everything else is a recommendation.

由此延伸出来的另一点是，物理并不在乎你的感受。

Another branch off of that is like, Physics doesn't care about your feelings.

所以他一遍又一遍地对那些人说：你怎么能这么快地评判别人、开除别人，或者整顿团队呢？

And so that is a line that he repeats over and over again with people who are like, How can you be, you know, so quick to judge people or to fire people or to, like, get the team?

他说，我们只有一个测试标准。

And he's like, we have we have one test.

我们只有一个必须达到的门槛。

We have one bar to clear.

火箭能飞起来吗？

Like, does the rocket fly?

我们能将多少有效载荷送入轨道？

And how much payload can we get to orbit?

我们只关注一个衡量指标，我会不惜一切代价推动我们达到这个目标，跨越这道门槛。

Like, there is one metric that we're optimizing for, and I will do what I have to do to march us towards that piece to clear that bar.

我的意思是，这就是阿斯伯格综合征真正带来优势的地方。

I mean, this is where the Asperger's comes out, you know, like in a truly advantaged way.

他说，渴望被喜欢是一种巨大的劣势，很多人都有这个问题。

Like he says being the need to be liked is like a huge disadvantage that so many people have.

而我没有这种倾向。

And I do not have that.

他直接说，我没有这个弱点。

He straight up says like, I do not have that weakness.

彼得·蒂尔对这一点有一个有趣的观察。

And Peter Thiel has an interesting observation about that.

他说，为什么这么多成功的科技创始人似乎都处于某种谱系上？

Was like, why are so many of the successful tech founders, like, seem to be somewhere on the spectrum?

为什么他们都具备这种在无视他人意见方面的生物学优势？

Like, why do they all have this sort of biological advantage in dismissing the opinions of others.

这说明了我们的社会什么样的问题？——竟然需要具备这种能力，才能成为创新者，打造新事物和伟大的组织？

What does that say about our society, that that's a skill that you need to have in order to be an innovator and build something new and build a great organization?

所以我想这引出了我接下来想问你的事情，也就是算法，这是算法的第一步。

So I guess that leads into the next thing I was going to ask you about, which is the algorithm, which is the first step of the algorithm.

算法是什么？

Algorithm is what?

问题的要求。

Question requirements.

他经常特别说，别把要求定得那么蠢。

Make your requirements less dumb is like specifically how he says it a lot.

要求太多了。

There are so many requirements.

我认为这确实是一个来之不易的第一步。

I think it's a really that was a hard earned first step.

是的，说说他当时是怎么做的。

Yeah, explain what he had

他是怎么学会的。

to go, like how he learned.

聪明的工程师最常见的错误，就是优化本不该存在的东西。

The most common error of a smart engineer is to optimize something that shouldn't exist.

需求的来源有很多，对吧？

And there are many places that requirements come from, right?

有时这些需求来自组织内部。

Sometimes they come from inside the organization.

有时它们来自某个团队。

Sometimes they came from a team.

有时它们来自法务部门。

Sometimes they came from legal.

有时它们来自设计合作伙伴，比如NASA。

Sometimes they came from a design partner, like in NASA.

比如，如果你要发射火箭，飞船要与空间站对接，他们对你的锁扣和各种连接方式都有各种要求。

Like, if you're gonna shoot up a rocket and the capsule of the vehicle is gonna, like, dock with the space station, they have all kinds of requirements about your latches and, how things go in.

如果你要从特定的发射台发射，就会有关于声音、时间或其他各种因素的要求。

If you're gonna launch from a specific pad, you've got requirements around sound or timing or all these things.

他说，有太多的钱、时间和精力浪费在试图满足那些根本就毫无意义的需求上。

And he he's like, so much money and time and effort is wasted trying to accommodate requirements that don't make any fucking sense in the first place.

所以首先要做的，就是缩短你的设计需求清单，以便获得尽可能大的发挥空间。

So the first thing to do is, like, shorten your list of design requirements so that you've got the maximum possible space to play in.

还有很多次，有很多故事，比如马克斯的文章里提到的那个真正的例子：要求必须从一家航空航天公司购买特定的Omni锁具，而且必须用这些材料制造，价格贵得离谱。

And there's so many times, many stories, and the latch is like a true one that's in Max's essay, which is like the requirement said, you got to buy this specific like Omni latch from an aerospace company made of these materials and it was like super expensive.

他们对这个要求提出了质疑，心想：只要它能完成这些功能，那就没问题。

And they sort of questioned that requirement and they were like, all well, fine as long as it like does this and this, it's fine.

于是他们从McMaster Carr买了现成的零部件，结果做得很好，价格却便宜了两个数量级。

And so they bought off the shelf parts from like McMaster Carr and they did the job just fine, which were like two orders of magnitude cheaper.

还有些时候，这些要求是法律或监管方面的。

There's also times when the requirement is legal or regulatory.

他会指派一个人去处理。

And he'll assign somebody.

他会说：去吧，申请这个许可证可能要花一年甚至一年半的时间。

He's like, go it'll take a year or a year and a half to like get this permit.

他会说：指派一个人去。

And he's like, assign somebody.

他会说：去那个办公室，坐在那里，直到你拿到许可，签了这张表。

He's like, go fly to this office and sit in this office until you get permission, like sign on this sheet.

回来。

Come back.

没带签好的表格就别回来，否则被开除。

Don't come back without a signed sheet or get fired.

这是一种质疑、取消或豁免该要求的方式，因为这根本毫无意义。他还提到，我们被迫遵守许多二十年前、五十年前的规定，这些规定完全不合理，也对消费者毫无帮助。

Like, that is a way to question the requirement or a dismissal or a waiver of this requirement because it doesn't make any And there's things that he's like, we are forced to comply with so many regulations from twenty years ago, fifty years ago, they just don't make any damn sense and they're not helping the consumer.

所以，让我们尽可能删除这些规定，以便以最低的成本设计出最优秀的产品。

So let's remove as many of them as we can so that we can design the best possible product for the cheapest possible price.

第二个是什么？

What's the second?

删除，删除，删除。

Delete, delete, delete.

简化。

Simplify.

关于这一点，经典的说法是：最好的部件是没有任何部件，最好的流程是没有任何流程。

There's a I mean, the canonical line about this is like the best part is no part, best process is no process.

我不确定具体数字，但在Model 3的研发过程中，合并零件或彻底取消零件的次数非常多。

And I don't know exactly the number, but the number of times combining parts has happened over the course of the Model three or eliminating them.

我觉得他刚起步时，零件数量超过了一万个。

Like, I think when he started, there were over 10,000 parts.

随着他不断合并零件，这个数字一直在稳步下降，因为每当有两个部件需要连接时，不仅需要连接这两个部件，可能还需要粘合、拧螺丝、上螺栓、铆接等等。

And that has come down steadily as he's combined parts because every time two things need to be attached, not only are there two things that need to be attached, but maybe they need to be glued, maybe they need be screwed, bolted, riveted, whatever.

金属之间的相互作用也需要被考虑。

The way the metals interact needs to happen.

它们必须真正地被组装起来。

They need to be actually assembled.

而且还会存在公差问题，可能影响其他部件的适配。

And then there's like a tolerance that could change whether other things fit.

额外的零件或额外的连接所带来的成本、变更和风险高得惊人。

It's like the number of expenses, changes and risks that come from additional parts or additional attachments is so, so, so high.

简洁性既能带来可靠性，也能降低成本。

Simplicity gets you both reliability and low cost.

就是这样。

That is it.

最好的部分就是没有零件。

The best part is no part.

简约同时带来了可靠性和低成本。

Simplicity delivers both reliability and low cost.

因此，在这两者之间，你必须尽可能彻底地删减。

And so between those two things, you have to delete as much as humanly possible.

所以他不遗余力地合并零件、简化设计。

And so he goes to great lengths to combine parts, simplify things.

他说，要极致地坚持删减和简化。

He says go ultra hardcore on deletion and simplification.

这就是你看到的那张关于Raptor的梗图，事物变得越来越简单。

That's how you see that, like, raptor meme, where things get more and more and more simple.

Model 3就是这样变得越来越简单、更可靠、生产成本更低、也更好。

And that's how the Model three has gotten simpler, more reliable, cheaper to produce, and better.

也许最著名的例子是他决定一体铸造整个前部和后部。

And maybe the most famous example of this is his decision to cast entire the front third and the back third.

这个想法来源于玩具车。

And he got that idea from toy cars.

玩具便宜、可靠，而且能快速规模化生产。

Like, toys are cheap, reliable, scale quickly.

他们是怎么做出来的？

How do they make them?

让我从这个行业中借鉴一个点子。

Like, let me steal an idea from this industry.

结果发现，从来没有人制造过这么大的铸造设备。

And it turned out that nobody had ever made casting machines that big.

好的，这一点很重要，因为他把这些想法结合在了一起。

Okay, so this is an important part, because then he combines these ideas together.

我们刚刚谈到了第一性原理思维，我认为他基本上是推广这种思维的人。

We just talked about first principles thinking, which, know, I think he was the one who basically popularized it.

现在人人都在谈论它，但真正去实践的人却寥寥无几。

Now everybody talks about it, very few people actually do it.

但在那本书里，我其实从我一集节目中剪了一个片段，因为我觉得特别有意思——他在特斯拉时，正摆弄着一个小小的玩具特斯拉。

But in that book, and I actually made a clip this from one of my episodes because I thought it was so interesting, where he's like at Tesla, he's just playing with like a little toy Tesla.

他还谈到从乐高中学到的东西，以及它需要多么精准。

And he talked about like learning from Legos and like how precise it has to be.

于是他们问：为什么我们不能直接铸造整个底盘呢？

And they're like, Why can't we just cast the entire, I think, underbody?

他回答说：因为没有那么大的铸造机。

And he was like, because there's no casting machine that's that big.

他说，他很确定，有多少家公司……

He said, Well, how many companies he's pretty sure.

有多少家公司制造铸造机？

How many companies make casting machines?

我想全世界大概只有六家左右。

I think there was like something like six in the world.

五家，他说，那我们去跟他谈谈吧。

Five, they He's like, Well, let's go talk to him about it.

五家说不，一家说也许可以。

Five said no, one said maybe.

埃隆说，我把那个‘也许’当成是同意了。

And Elon goes, I took that maybe as a yes.

这大大简化了生产流程。

And it greatly simplified, like, the production.

书中还有另一个故事，我觉得他也在几个播客里提到过，当时他们在解决特斯拉的生产困境时，不得不 literally 拆掉一堵墙。

There's another story from the books, and I think he talked about this on a couple podcasts, too, where it's like the They had a When they were trying to solve production hell of Tesla, they had to literally knock a wall.

他们淘汰了太多机器人，处置速度太快，于是就在工厂侧面砸开了一面大墙，直接扔掉了——我忘了具体多少，几百台机器人。

They were throwing out so many robots disposing them so fast, they knocked a huge wall in the side of the factory, just threw out like, I forgot how many, hundreds of robots.

是的。

Yeah.

因为他发现了一些奇怪的情况，比如，这个机器人抓起一个零件，然后交给另一个机器人。

Because he found weird shit like, Hey, this one robot picks up this piece, hands it to another robot.

那个机器人然后转过身把零件交给另一个机器人。

That robot then turns around and hands it to another one.

他问：为什么这个机器人不直接把零件交给那个机器人呢？

He's like, Why don't this robot just hand it to that robot?

为什么中间要多出这么一个人？

Why is this guy in the middle?

这增加了复杂性。

And it adds complexity.

他说：直接把它删掉。

He's like, Just rip it out of there.

是的。

Yeah.

这在生产系统里就是这样。

And that's in the production system.

是的。

Yeah.

意思是，自动化是算法的最后一步。

Mean, this is why automate is the last step of the algorithm.

因为他以前总是先这么做，对吧？

Because he used to do that first, right?

是的。

Yeah.

他最初对特斯拉的设想是，我们要自动化一切。

His original vision for Tesla is like, we're going automate everything.

这将是一个纯粹的机器人，你知道的，全自主的生产线。

This is going to be like a pure robot, you know, dreadnought autonomous line.

他说，这是一个巨大的错误。

And he's like, that was a huge mistake.

我们还没有把它完善。

We didn't have it refined.

我们的设计还没有完善。

Like, we didn't have the design refined.

我们没有完善的制造设计。

We didn't have the manufacturing design.

我们没有质疑过需求。

We hadn't questioned the requirements.

我们没有删除足够的零部件。

We hadn't deleted enough parts.

在完成所有这些其他步骤之前，我们不应该进行自动化。

We should not have automated until all of these other steps had been gone through.

他说，他多次完全颠倒了这些步骤的顺序。

And he said there's many times where he did every step in exactly reverse order.

有些东西先是被自动化了，然后试图加快速度，接着优化，最后却因为没人质疑需求而被删除了。

Something was like automated and then tried to be done faster and then optimized and then ended up getting deleted because nobody had questioned the requirements.

这就是为什么我们要一遍又一遍地按这个顺序来做这些事。

That's why we do these things in this order over and over and over and over again.

还有一个关于星链项目的故事，当时简直一团糟，对吧？

And oh, there's a great story of taking Starlink was like a mess, right?

它贵了十倍，而他们建造的数量只有所需量的十分之一。

It was 10x too expensive and they were building one tenth of how many they needed.

所以，成功程度差了两个数量级。

So, like, two orders of magnitude off success.

他说道：我受够了。

He's like, I've had it.

我要解决这个问题。

I'm fixing this.

现在这成了瓶颈。

This is now the bottleneck.

他找来了马克·乔根森，他是SpaceX早期招聘的顶尖人才之一。

He grabs Mark Jorgenson, which is one of his, like, super talented early hires from SpaceX.

这是一位火箭科学家。

This is a rocket scientist.

他从未从事过卫星相关的工作。

This is not a guy who's ever worked on satellites.

他召集了一支他信任的工程师团队。

He grabs a team that of engineers that he trusts.

他们飞往西雅图。

They fly up to Seattle.

他们解雇了整个星舰领导团队和星链领导团队。

They fire the entire Starship leadership our Starlink leadership team.

然后他们进入作战室，我认为这是一种被低估的策略。

And they sit down in a war room, which I think is an underrated like tactic.

他们开始运行算法。

And they start running the algorithm.

卫星设计的第一性原理是什么？

What is the first principles of like satellite design?

我们能把这个东西做得多简单？

How simple can we make this thing?

为什么？

Why?

为什么？

Why?

为什么？

Why?

为什么？

Why?

为什么会有这个东西？

Why does this exist?

为什么这两样东西差距这么大？

Why are these two things so far apart?

为什么我们需要这么多能量？

Why do we need this much energy?

为什么我们需要这个制造流程？

Why do we need this manufacturing process?

在几个月的时间里，他们实现了两个数量级的飞跃。

And over the course of like a few months, they make this two order of magnitude leap.

比如，Co.

Like, Co.

他说，这些从未接触过这种设计的人，仅仅通过应用该算法，并以狂热的紧迫感追求极高的设计标准，就创造出了这款产品——如果它是一个独立企业，如今价值可达数百亿美元。

Said, these people who had never encountered this design before, but just by applying the algorithm and working with maniacal urgency towards this extremely high design bar, they created this product that's now, you know, if it was a standalone business, would be worth tens of billions of dollars.

Deel 是全球在构建跨国招聘基础设施方面最出色的公司。

Deel is the best company in the world at building infrastructure for global hiring.

Deel 将帮助您的企业在全球任何地方招聘、支付和管理员工。

Deel will help your business hire, pay, and manage any worker anywhere in the world.

Deel 是一个平台，可覆盖全球150个国家的薪资、人力资源、福利和设备管理。

Deel is one platform for payroll, HR, benefits, and device management across a 150 countries.

Deel 提供您在单一AI原生平台上运行高效全球团队所需的一切。

Deel gets you everything you need to run a high performing global workforce on a single AI native platform.

从第一份录用通知到最终离职流程。

From the first offer to final offboarding.

Deel 处理所有复杂事务，让您专注于业务本身。

Deal handles the complexity so you can stay focused on your business.

世界上最好的创始人和运营者有一个共同点。

The best founders and operators in the world have one thing in common.

他们尽可能掌控自己的业务。

They control as much of their business as possible.

Deel 的创始人正是这样做的。

The founders of Deel do exactly this.

当你使用 Deel 时，你并不是在使用第三方薪资处理服务或混乱的本地供应商网络。

When you use Deel, you aren't using a third party payroll processor or a messy network of in country providers.

Deel 自主构建并拥有自己的基础设施。

Deel built and owns their own rails.

这意味着更快的速度、更好的服务和完全的责任制。

That means faster speed, better service, and total accountability.

Deel 已经构建了能够采取行动的 AI 同事。

Deel has built AI teammates that take action.

大多数人力资源 AI 仅止步于聊天。

Most HR AI stops at chat.

Deel 将人工智能嵌入到工作流程中，使其能够基于经过合规审查的知识，执行招聘、薪酬、人员调动、IT 支持和报告等实际工作。

Deel puts AI inside the workflows themselves so it can run real work across hiring, payroll, mobility, IT support, reporting grounded in compliance vetted knowledge.

Deel 的人工智能会主动采取行动，帮助团队在不增加人员的情况下更快推进工作。

Deal AI takes action to help teams move faster without adding headcount.

Deel 获得了超过四万家企业的信赖。

Deal is trusted by over 40,000 businesses.

立即访问 deel.com，了解他们如何帮助您的企业。

Learn how they can help your business today by going to deel.com.

那就是 deel.com。

That is deel.com.

我制作了这期名为《埃隆是如何工作的》的《创始人》节目。

I did this episode of Founders called How Elon Works.

我想在简介里，甚至在开场时，我就说过：听好了。

And I think in the description, maybe even in the intro, I was like, listen.

我通常已经读过四百本这种该死的传记了。

I can usually I've read 400 of these fucking biographies.

通常来说，我都能找到一个与当今企业家相似的历史人物，可能五十年前、一百年前就有这样一个人。

Like, I can usually find a historical analogy to an entrepreneur living today where like that person was very similar to this person maybe fifty years ago, one hundred years ago.

我认为埃隆是独一无二的，而且他说过，他是历史上唯一一个活着或已故的这样的人。

I think Elon's singular, and I think he said he was like singular living or dead.

我真的找不到一个历史上的类比，但如果是这样，他反复提到——这是直接引述，也是那本书中对埃隆话语的转述——他说：‘我反复讲这个算法，讲了太多遍，以至于开会时人们都能预测我接下来会说什么了。’

I can't really find a historical analogy to But it's like, if that's the case, and he repeats, he says, this is a direct quote, it's a paraphrase of a quote from Elon in that book, where he's just like, I would say the algorithm so much, and I repeat it so many times that people in the meeting would start to be able to predict what's coming out of my goddamn mouth.

所以，我们为什么不认为，这个人在创业史上是独一无二的，而且他一遍又一遍地重复这些话呢？

So why don't we think about this guy doesn't have he's a singular character in the history of entrepreneurship, and he repeats this over and over again.

也许我们应该花更多时间，反复思考这个问题，因为这个故事就是一个绝佳的例证。

Maybe we should spend a little bit more time, and we should think about this over and over again, to the point that was a great illustration of the story.

我认为这里的格言是：重复具有说服力。

I think the maxim here is repetition is persuasive.

就像你必须反复强调一件事，这件事在许多创始人身上都反复出现。

It's like you have to repeat One thing is reoccurring throughout a lot of these founders.

他们需要识别出少数几条原则。

They to identify a handful of principles.

如果我们翻遍所有笔记、所有书籍，以及我们面前的所有资料，也不会有25个不同的想法。

If we go through all the notes, all the books, everything we have in front of us, it's not 25 ideas.

他只有几个核心原则，反复不断地使用它们。

He's got a handful of principles that he uses over and over and over again.

我特别喜欢这个想法，他构建的这种模式几乎像另一种形式的网络迷因。

And I love this idea that he built like this, almost like another form of a meme.

这就像一个操作系统，包含四个步骤，我们可以将其应用于太空、汽车、隧道等各种领域。让我们来看第三步。

It's like this operating system where there's these four steps, and we can apply it in space, in cars, in tunnels, and everything Let's go to step number three.

第三步是简化和优化，我认为这正是大多数工程师平时就会做的标准操作。

Step number three is simplify and optimize, which is I think that's what most engineers just, like, do as a standard.

但你必须在确信这个东西确实存在，并且已经把需求尽可能清晰、简洁地明确之后，才能进行这一步。

But that is the thing that, like, you should only do after you know for sure that the thing exists and you know for sure you've got your requirements as clean and short as you possibly can.

关于需求的另一件非常重要的事是，每一个需求都必须明确归属于某一个具体的人。

The other thing that's actually really important about the requirements is that a single individual's name has to be attached to every one of them.

他非常强调个人直接负责。

He's huge on direct individual accountability.

所以根本不存在什么来自法务部门的要求。

And so there's no such thing as like a requirement from legal.

而是来自法务部门的JAM提出的要求，这样你就可以去问她：这个要求为什么存在？

There's like a requirement from JAM in legal so that you can go ask her and be like, why does this requirement exist?

请非常详细地解释一下。

Please explain it very thoroughly.

我们能把它删掉吗？

Can we throw it out?

诸如此类。

Etcetera.

对吧？

Right?

很多时候，你会发现一些要求是由两年前离职的实习生，或者一个没人真正认同该要求的部门制定的。

So many times you found requirements that were created by an intern who left the company two years ago or a department where nobody in the department actually agrees with the requirement.

所以一旦你搞清楚了，接下来就是很常规的工程工作了：让我们分析一下权衡取舍，优化这个东西。

So once you get there, it's like the very normal engineering work of like, let's figure out the trade offs, let's optimize this thing.

是的，特斯拉曾有一个很有趣的故事，当时电池包上面有一层东西。

Yeah, there was a great story from Tesla where you had the battery pack and then it had something like a layer on top of it.

这一层是用来隔音的，但特斯拉的一部分人认为它是用于防火，另一部分人则认为它是用于隔音。

This was for sound, and one part of Tesla thought it was for sound and the other thought it was for, like, fire prevention.

我记的是不是对的？

Am I remembering this correctly?

对。

Yeah.

防火团队以为那是用来隔音的，而隔音团队则以为那是用来防火的。

Like, the fire prevention team thought it was for sound baffling, and the sound team thought it was for fire protection.

一旦明确了这一点，他们实际上就做了测试。

And, like, once that was established, they actually they tested it.

他们在车里放了一个麦克风，结果根本听不出区别。

They put a mic in the car, like, nobody could tell the difference.

他们说：删掉吧。

They're like, delete.

好的。

Okay.

所以但是

So but

在那个特定时期，它是个瓶颈。

it was the bottleneck at that particular time.

但再说一遍，这个部门这么说，那个部门那么说。

But let let like, again, it's like this department said this and this department said that.

他们意见不一致。

They're not agreeing.

但我觉得不能跳过这一部分，因为我认为这是最重要的部分，埃隆的解决方案非常简单。

But then I think not to skip over this part because I think it's the most important part, Elon's solution was very simple.

就是说，它是否减少了？

It's like, well, does it reduce?

在车内，如果这个东西不存在，会不会更吵？

Inside the cabin, is it louder if this thing's not here?

我们该怎么弄清楚呢？

How are we going figure it out?

好吧，干脆直接放个该死的麦克风，然后测试一下。

Well, let's just put a fucking microphone and we'll test it.

我们来测试一下。

We'll test it.

六小时后，这个部件就被删掉了。

And six hours later, the part is deleted.

就像这个一直拖慢进度的东西，他们从生产线上撤掉了更多机器人，结果速度就上去了。

Like this whole thing that was slowing down, you know, they pulled more robots off the line and it sped up.

这提高了整个生产线的产能。

And that increased the throughput of the entire production line.

然后，这是一个价值数百万、也许是数千万美元的决策。

Then it was millions, maybe tens of millions of dollars worth of decision.

他说，有时候一场仅仅半小时的会议，就能为特斯拉的企业价值增加一亿美元。

Like, he says there are times when a single half hour meeting has added a $100,000,000 to the enterprise value of Tesla.

每一分钟的思考都价值百万美元。

Every single minute of thinking is like worth a million dollars.

高质量的一分钟思考。

A high quality minute of thinking.

这些工具就是他在那些分钟里做的事情，真正产生了这样的影响，正如你所说，他是我们这一代最杰出的创业者。

And these tools are the things that he does in those minutes that actually have that kind of effect, which is, to your point of like, this is a singular greatest entrepreneur of our generation.

这是他最常提到的事情。

This is the thing that he talks about most often.

也许这很有用。

Maybe it's useful.

我大概有六个月没做那一集了，我重新听了一遍，为的是准备和你的这次对话。

Well, I hadn't made that episode in probably like six months, and I was listening to it again to prepare for this conversation with you.

我觉得我应该每个月都听一遍。

I'm like, I should listen to this every month.

就一个小时半小时而已。

It's just like, it's an hour and a half.

你可以用1.5倍速听。

You can listen at 1.5.

如果你以1.5倍速播放，也就一个小时左右。

It's an hour if you're at 1.5 XP or whatever it is.

我只是觉得，哦，我忘记了一些这些想法。

It's just like, Oh, I forgot some of these ideas.

我认为这就是重复如此重要的原因。

And I think that's where the repetition is so important.

是的。

Yeah.

我的意思是，这正是我设计所有书籍时都希望它们能被反复阅读的原因。

I mean, that's why I really design all my books to be read and reread.

我希望它们能成为手册和参考指南。

Like, I want them to be handbooks and reference guides.

我特别喜欢看到那种被翻得破破烂烂的书，被人每天带在背包里，几年来反复翻阅，书角卷起、贴满便签、每一页都画满了重点。

I love when I see, like, just a beat to shit copy that's been in and out of somebody's backpack, like, every day for years and is like dog eared and flagged and all the highlight on every page.

这些想法，重复并不会削弱其力量，重复才是有说服力的。

Like, these ideas, repetition doesn't spoil the prayer, repetition is persuasive.

你必须努力把这些想法植入脑海，这样当你需要时，它们就在那里——当你面对这些决策时，能提出正确的问题或拥有正确的直觉。

Like, you have to work to install these ideas in your head so that they're there when you need them, that you ask the right question or have the right instinct as an operator when you're confronted with one of these decisions.

你还能想到其他来自算法第三步的有趣故事吗？

Are there any other interesting stories that you can think of that come from the third step of the algorithm?

说实话，第三步是最无趣的。

I think the third step is the least interesting, honestly.

提问要求和删除可能占了80%的工作量。

Question requirements and delete are probably 80% of the lift.

我认为简化与优化、加速与自动化，才是人们自然会优先想到的。

I think the simplify and optimize, the accelerate and the automate are the things that people leap to naturally.

这个算法真正设计的目的是让他们先磨利斧头，再开始砍树，或者确保你砍的是正确的那棵树。

And the algorithm is really designed to get them to, you know, sharpen the axe before they start chopping down the tree or be sure you're chopping down the right tree.

确保砍树本身根本就是正确的下一步。

Be sure chopping down the tree is the right next step at all anyway.

算法的第四步是什么？

What's step four in the algorithm?

加速。

Accelerate.

更快一点。

Go faster.

不管你现在的速度有多快，你都能再快一点。

However fast you're going, you can go faster.

还有别的吗？

Is there anything else to it?

没有。

Not

真的没有。

really.

所以就是加快速度。

So just go faster.

他有句话，我们该怎么说来着？

He has that line where we what do you say?

一种狂热的紧迫感将成为我们的运作原则。

A maniacal sense of urgency is gonna be our operating principle.

是的。

Yeah.

但举个例子，我可能会想到隧道掘进机。

But, like, the example here I would use is probably going to the tunneling machines.

他说，如果你想加快隧道施工，目前那些在先进隧道施工中运行的机器，远未达到它们的功率或热能极限。

He's like, if you want to accelerate tunnel like, the machines that are currently tunneling, even in state of the art tunneling, are nowhere near their power or thermal limits.

所以，干脆把它们调到最大速度。

So just, like, crank them up.

让它们以最快的速度运行。

Go put make them go as fast as they can.

当一切真正实现自动化时，有趣的是，机器人能移动的速度限制，远高于人类能移动的速度。

Like, can the ro the interesting thing about when everything is actually automated is, like, how fast can the robots move is a way higher limit on that than how fast can humans move.

所以如果你真的加快了所有流程，他一直在谈AI，比如，如果你能移除人为因素并实现全面自动化，就能达到一种超人类级别的速度。

And so if you actually do accelerate everything, he's been talking about this a bunch with AI, like, if you can remove the human element and then you can automate everything, you can do it to a degree that is, like, superhuman level of speed.

你在书里对这一点的描述非常精彩。

You have a great way to describe this in the book.

我直接念一下书中埃隆对第四步的描述，我非常喜欢这段话。

I'm just going to read this is Elon's description of step four from the book, which I love.

他说，在第四步，也就是只有在前面几步都完成之后，才能加速周期时间。

And so he's talking, he's like, then and only that in step four, accelerate cycle time.

一旦你已经朝着正确的方向高效前进，但仍然觉得太慢了。

Once you're moving in the right direction and moving efficiently, you're moving too slow.

再快一点。

Go faster.

你总能让事情变得更快。

You can always make things go faster.

但在你完成前三件事之前，不要急于加快速度。

But do not go faster until you've worked on the other three things first.

我错误地花了大量时间加速一些流程，后来才意识到这些流程本该被删除。

I mistakenly spent a lot of time accelerating processes that I later realized should have been deleted.

加快这个过程真是太荒谬了。

Speeding this is hilarious.

加速本不该存在的东西是荒唐的。

Speeding up something that shouldn't exist is absurd.

如果你正在挖自己的坟墓，别挖得更快。

If you're digging your grave, don't dig it faster.

别再挖了。

Stop digging.

是的。

Yeah.

这在书中关于组织设计的部分，我觉得他把这套算法应用到自己的公司里这一点被严重低估了。

This is in the designing the organization section of the book, and I feel like it is really underrated that he applies the algorithm to his own companies.

比如，他尽量缩短设计师、工程师和制造工程师之间的距离，并尽可能地精简公司里的各种事务。

Like, he minimizes the distance between the designers, the engineers, the manufacturing engineers, and gets rid of as much stuff in his companies as he possibly can.

他让公司运作得更快。

He makes the companies go faster.

他会质疑公司的需求和结构本身。

He questions the requirements of the companies and the structure themselves.

他还谈到如何通过产品看出组织中的缺陷。

And you can he talks about how you can see the flaws in the organization in the product.

比如，电池团队在为电池设计一个外壳，而车体的底盘正好在它上方，他们还是会为电池加上一个盖子，因为从设计一个完整的电池来看，有盖子似乎更合理。

So if there's a battery team building a box for the battery, but a chassis in the car right above it, like, they'll put a lid on the battery because it makes sense to design a complete battery with a lid.

但车体就在电池上方，所以电池外壳其实根本不需要盖子。

There's a car on top of it, so you don't actually need a lid on the battery case.

最终你得到的是一个盒子套着另一个盒子，而最高效的做法是让这些团队之间有足够的互动，从而避免重复的零件、职责或材料。

You end up with, a box in a box where the most efficient thing to do is, like, have enough interactions between all those people that you are not duplicating parts or responsibilities or materials.

因为如果你允许各个独立部门做出过多的优化决策，他们只会优化自己小团队的利益，而不会考虑这个部门在整个大系统中的角色。

Because if you allow individual pockets to make too many optimization decisions, they'll optimize for their own tiny unit, but not how the unit exists in part of the broader system.

你会看到一堆不兼容的金属材料需要拼接，或者一堆荒谬的设计决策，导致整个产品无法达到最优。

You'll have like a bunch of disparate metals that have to come together or a bunch of insane design decisions that don't make the full system optimal as a whole product.

是的，他谈到了与此相反的情况。

Yeah, he talks about the opposite of that.

我认为他称之为象牙塔式工程。

I think he calls it ivory tower engineering.

对。

Yeah.

他说，我会设计一个产品或流程。

He's like, I'm going to design a product or process.

我会把它扔过墙去，你们负责工程或制造。

I'm going to throw it over the wall and you can engineer or manufacture it.

他说，不，工程师和制造商必须能够去问设计师：你为什么这么做？

And he's like, No, you have to be able to the engineer and the manufacturer should be able to call our designer, like, Why did you do it this way?

是的。

Yeah.

回到迭代速度和你可以纠正和识别的失败次数，比如推进猛禽发动机的生产。

Going back to the iteration speed and the number of failures that you can correct and identify, You know, moving the Raptor production.

他让整个团队把办公桌搬到了猛禽发动机的生产线上。

He had their whole team move their desks to the Raptor production line.

这样，当他们坐在那里设计或工作时，无论做什么，都能亲眼看到机器和发动机就在他们旁边制造出来。

So as they're sitting there designing or working on it, whatever they're working on, they're watching the machine, the engine, get manufactured right next to them.

他们之间不断交流。

And they're dialoguing.

这种反馈机制的带宽，连接着真正处理实物的人和设计产品的人，高得惊人。

This is like the feedback, the bandwidth between the people actually moving the atoms and the people designing the product are super, super, super high.

我的意思是，我觉得他提到传统航空航天业就像是设计一个文件，然后拆分出去，交给分包商，这些分包商再转给更下一级的分包商。

I mean, I think he talks about traditional aerospace is like designing a file that then gets parted out, that goes to subcontractors, and those subcontractors delegate it to subcontractors.

你得往下找五家公司，才能找到真正切割金属或接触材料的人。

And you have to go five companies down before you find a guy who's actually like cutting metal or touching the material.

然后信息又一路反馈回去，每一层都在增加复杂性、混乱、管理成本和利润。

And it kind of goes all the way back up and everybody's adding like complexity and confusion and overhead and cost and profit.

这也是SpaceX能够做到某些事情的部分原因。

And that's part of the reason why SpaceX was able to do something.

效率和成本效益高了几个数量级。

Was like orders of magnitude more efficient and cost effective.

而且这还没算上他们设计迭代速度极快，不断去除所有不必要的部分。

And that's before you even get to the fact that they're like iterating so quickly on the design that they're removing all this unnecessary stuff.

好的。

All right.

我们进入第五步。

Let's go to step five.

我想我误以为只有四个步骤。

I think I mistakenly said there's four steps.

我已经做了十期相关内容，现在早该注意到了。

I've done 10 episodes on this, I should notice by now.

所以，我们来看第五步。

So let's go to step five.

第五步是自动化。

Step five is automate.

这是终极目标，但你必须确保自动化的是正确的事情，因为自动化本身——创建自动化的过程——成本高昂且难以调整。

It is the holy grail, but you want to be sure you're automating the right thing because automation itself, the process of creating the automation is expensive and difficult to adapt.

所以，是的，事情应该被自动化，但只有在你彻底执行了算法的其他所有步骤之后，才应该进行自动化。

So, yes, things should be automated, but then it should only be automated after you've rigorously applied the other steps of the algorithm all the way through.

你刚刚提到一件事，我认为这贯穿了他整个职业生涯的一个重要主题：他总是想掌控一切。

You just mentioned something I think is a huge, like, theme that runs throughout, like, his entire career that he wants to control everything always.

是的。

Yes.

你能谈谈为什么这一点如此重要吗？

Can you talk about, like, why that is so important?

因为在特斯拉早期，他有过一段艰难的经历。

Well, he had a tough experience in the early days of Tesla.

我觉得，是关于分享责任或对公司权力的控制。

I think, like, sharing responsibility or power over this company.

特斯拉的起步阶段相当混乱。

And the beginning of Tesla was kind of a messy experience.

他是主要投资者和董事长，但并不是真正的首席执行官。

Like, he was the main investor and chairman, but he wasn't actually CEO.

但他花了大量时间深度参与产品，努力推动实现他对于电动汽车的愿景标准，但长期来看这并不可行，公司正走向一条艰难的道路，未能取得所需的进步。

But he spent so much time, like, getting deeply involved in the product and, like, trying to drive a good outcome there that reached the bar that he had for his vision for electric cars that it just didn't really work in the long term and that the company was heading down this tough path and wasn't really making the kind of progress that it needed to.

他谈到自己最初不愿出任特斯拉首席执行官的犹豫。

He talks about his reluctance to step in as CEO of Tesla.

他说：如果我不这么做，公司就会倒闭。

He's like, If I didn't do that, the company would have died.

我提供了大部分资金。

I provided most of the money.

我在产品开发过程中已经尽我所能深度参与了。

I've been as involved as you could be in creating the actual contributing to the product.

但最终，他觉得，如果这事想成功，我就必须掌控全局，全身心投入。

But eventually, he's like, if this is going to work, I've gotta take control of it, and I've gotta go all in on it.

是的。

Yeah.

他谈到了试图使用莲花汽车底盘这个错误，因为最初他们确实用了莲花。

He talked about the mistake of trying to maybe because originally, they used, like, the Lotus.

对吧？

Right?

是的。

Yeah.

我认为他说的是，我们应该从一张白纸开始设计。

The Lotus chassis, I think is what He's it like, we should just design this from a blank sheet of paper.

如果你要做一些全新的东西，从一张白纸开始设计总是更好的。

It's always better if you're gonna do something new, like, just design from a blank sheet of paper.

这真是一个很好的教训，实际上也是他的一大主题：追求技术的强形式。

It is a really good lesson and actually a big theme of his is like go for the strong form of the technology.

不要试图改造以前的东西。

Don't try to adapt what came before.

要像从第一性原理出发去思考，打造真正超越以往的科技突破，彻底摆脱过去的限制——不只是过去的 Designs，还包括整个供应链。

Try to like think in limits, go to first principles, and build like a true leap in technology that totally leaves behind the constraints of the past, not just the past sort of designs, but the past entire supply chain.

即使你依赖行业现有的供应商，要想像从一张白纸开始那样实现彻底的创新，也会非常困难——那时你会质疑每一个需求，从最理想的产品形态这一柏拉图式极限出发。

Even if you were counting on the existing suppliers in the industry, it's going be really tough to fully innovate in the way that is possible if you start with a truly clean sheet of paper, you question every requirement, you start with the platonic limit of what is the maximum ideal possible product.

然后从那里倒推回来。

Then work backwards from that.

我们能多接近这个理想状态？

How close can we get to that?

不要从类比开始，比如：我们怎么才能比现有产品好10%？

Not start with an analogy of like, how can we be 10% better than what exists in the field?

我们怎么才能好20%？

How can we be 20% better?

完美是什么样子？

What does perfection look like?

最理想的状态是什么？我们能多接近它？

What is the absolute ideal and how close can we get to that?

如果有人告诉你做不到，就反问：为什么不能？

If anybody gives you a reason why you can't do that, start asking why.

开始深入挖掘，质疑各种需求，挑战这些假设。

Start drilling down, start questioning requirements, start pushing on those assumptions.

不过，我认为强调这一点很重要，因为如果你去阅读并研究精神分裂症的历史，就会发现所有线索都指向同一个方向。

Well, I think it's important to like hammer on this point though, because it's obvious if you read, like if you just study history of schizophrenia, like they all line up in the same place.

就像尽可能实现垂直整合。

It's like as vertically integrated as possible.

他们会掌控自己业务的方方面面，包括原材料等等。

They're going to control as much of their business, their raw materials and everything.

我不得不注意到的是，这正是汽车工业起步时的样子。

And what I couldn't help but notice is this is how the car industry started.

比如亨利·福特，他们几乎自己生产所有零部件。

Like Henry Ford, they made almost all their parts themselves.

当然，那时候都是手工制造。

Obviously, they were making it by hand.

在他找到大规模生产汽车的方法之前，这正是他的核心理念。

This is before he figured out how to mass produce a car, and that was like his main idea.

但那不是汽车供应链。

But it wasn't those auto supply chain.

对。

No.

你开始这么做，然后你就掌握了控制权。

You start doing it and then you have control.

他也极度痴迷于控制。

He's also obsessed with control.

他最终在1919年买断了投资者的股份，我想是这样。

He wound up buying out his investors, I think, 1919.

到1919年，亨利·福特拥有福特汽车公司100%的股份。

Henry Ford owned 100 by 1919, he owned 100 of Ford Motor Company.

这相当于你现在拥有一家价值200亿到400亿美元的公司100%的股权。

Was just like the equivalent of you owning 100% of like a $20 to $40,000,000,000 company today.

他如此痴迷于控制。

He was so obsessed with control.

他买了一条铁路。

He bought a railroad.

他真的买了一条铁路。

Like, he bought a railroad.

我不但想控制Model T和我所有汽车的每一个设计和零部件，还想确保这些零部件能送到我手上，我要掌控它们运输的整个物流过程。

It's like, not only do I want to control all, every single design and part that goes into the Model T and all of my cars, it's like, I want make sure that all those parts get to me, I want to control the logistics of them getting to me.

是的。

Yeah.

埃隆的特别之处在于，他完全背离了当时美国汽车行业的趋势——那种层层外包的模式。

And what's fascinating about Elon is like he just went and like bucked the trend of the American automobile industry at that time, which was like subcontractor, subcontractor, subcontractor.

他一遍又一遍地提到这一点。

He talks about this over and over again.

他们连自己的东西都不自己生产。

Like, they're just not even making their own shit.

对。

Yep.

他说，我要回到过去那种做法。

He's like, I'm just going to go back and do it the way it used to be done.

但我认为这之所以如此重要，不仅是因为你掌握了控制权，还因为这样才能打造出成本高效的產品。

But I think the reason that it's so important is, like, not only do you have control, but it's, like, how you make a cost efficient product.

因为如果你层层外包，想想他最初做计算时，还没提出‘傻瓜指数’这个说法，这个我们以后可以聊。

Because if you're having subcontractor after subcontractor, think when he did the initial calculation, he wasn't calling it the idiot index, which we could talk about.

但在当时研究火箭时，他意识到火箭成本的98%并不来自材料本身。

But at the time with the rockets, he realized that like 98% of the cost of a rocket was not from what it was made out of.

他说：‘钱到底花到哪儿去了？’

He's like, Where the hell is the money going?

他发现，钱是经过一个承包商，这个承包商又雇了另一个分包商，那个分包商再雇了另一个分包商，再雇一个分包商，最后才到真正动手弯金属、制造零件的人手上。

And he was like, Oh, it's going from this contractor who's This subcontractor actually hired another subcontractor who hired another subcontractor who hired another subcontractor before I get to the actual person, like you said, bending the metal or making the thing.

这背后其实有一个非常重要的深层原因，那就是他有着完全不同的愿景和激励机制。

There's actually a very important, like, meta to that, which is he has a completely different vision and set of incentives.

如今整个航空航天行业基本上都是依靠政府的成本加成合同来维持的。

The whole aerospace industry today, they've basically been supported by cost plus contracts from the government.

他们的动机是什么？

What is their incentive?

他们的动机就是尽可能多地花钱，因为他们能拿到预算的百分比。

Their incentive is to spend as much fucking money as they can because they make a percentage of the budget.

他们实际上并不被激励去成功。

They're not actually incentivized to succeed.

他们没有得到良好的激励，也许他们被激励去成功，但并不是被激励去经济上成功。

They're not incentivized well, maybe they're incentivized to succeed, but they're not incentivized to succeed economically.

他们也不一定被激励去按时完成原定期限。

And they're not necessarily incentivized to succeed on the original deadline.

埃隆的做法是去找真正发放这些合同的人，说：我们想要一个以成果为导向的合同。

What Elon did is went to the people actually awarding these contracts and said, like, we want an outcome based contract.

我希望被激励去降低成本。

I want to be incentivized to drive down the cost.

我希望被激励去按时完成期限。

I want to be incentivized to hit the deadlines.

他想要在低成本上竞争。

Like, he wanted to compete on low cost.

尽管人们总在谈论他做了前所未有的事情，但他所有这些前所未有的成就，都是因为他在降低成本。

Like, for all the talk that happens about him doing things that have never been done before, he is all the things that he's doing that have never been done before are because he drives cost down.

而他降低成本，是因为他的愿景不是赚尽可能多的钱。

And he drives cost down because the vision is not make as much money as I can.

他的目标不是创办一家利润尽可能高的新航天公司。

It wasn't start a new aerospace company that's as profitable as possible.

他的目标是让我们抵达火星，这意味着要把把有效载荷送离地球的成本降到最低。

It was get us to Mars, which means drive down the cost of getting a payload off Earth as low as possible.

这意味着我希望被激励去打造一个成本尽可能低的系统。

Which means I want to be incentivized to build a system that is as low cost as possible.

一旦他做到了这一点，就开始赢得所有这些合同，赢得这些真正以成果为导向的项目，而这些项目，顺便说一句，对政府和纳税人来说都更好。

And once he did that, he did start winning all these contracts and winning these, like actual outcome based things, which by the way, better for the government, better for taxpayers.

敢于站出来说，我知道我可以拿X钱去做平庸的工作，或者只稍微好一点的工作，这是一件了不起的事。

That is a heroic thing to go in and say, I know I could get paid X to do mediocre work or to do marginally better work.

但我真的想承担风险。

But I actually want to take the risk.

我想烧掉退路。

I want to burn the boats.

我想背水一战。

I want my back to the wall.

我想明确知道，我们必须跨越这个前所未有的门槛。

I want to know that we have to clear this threshold that nobody has ever crossed before.

我想告诉我的团队，因为这才是我们真正实现这一目标的方式——这个目标让所有人都觉得我疯了，那就是带我们去火星，实现人类工程系统效率上巨大的数量级飞跃。

And I want to go tell my team that because that's how we're actually going to achieve this objective that everybody thinks I'm crazy for setting, which is get us to Mars, which is make this very massive orders of magnitude leap in the efficiency of like a human engineering system.

这是一种完全不同的方法，这可以追溯到他的科幻梦想和以使命为导向的本性。

It is such a different approach and that goes all the way back to his like his sci fi dream and his purpose driven nature.

特斯拉也是如此，他并没有说，我想开一家盈利的电动汽车公司。

It's the same thing with Tesla, that he didn't go out and say, I want to make a profitable electric car company.

他的想法是，我希望尽可能多的人驾驶电动汽车，因为这才是解决气候变化的方式。

He's like, I want as many human beings as possible driving electric cars because that's how we're going to solve climate change.

这就是我们如何在这上面留下印记的方式。

That's how we're going to put a dent in this.

所以他的所有决策都不是如何在制造电动车的同时最大化利润？

And so all of his decisions are not how do I maximize profits while making electric cars?

而是如何让尽可能多的人驾驶电动车？

It is how can I get as many people as possible driving electric cars?

它们必须尽可能便宜。

They have to be as cheap as possible.

我必须把成本降下来。

I have to drive that down.

我需要简化产品。

I need to simplify the product.

我需要实现大规模产量，使制造规模能够支撑一个极低的价格点。

I need to get a massive volume so that the scale of the manufacturing supports a really, really, really low price point.

因此，他始终从长远视角出发，做出这些基于原则的决策。

And so he just keeps making these like principle driven decisions from a very long term view.

他的视野如此宏观，以至于实际上很难让人理解，但他不断做出一些看似疯狂、愚蠢、富有远见或适得其反的决定，却并不违背行业的主流思维。

Like his view is so zoomed out that it's actually really difficult to relate to, and he keeps doing these things that seem crazy or stupid or visionary or counterproductive, but isn't counter to the prevailing wisdom of the industry.

但这并不是因为别人这么做，我就也要这么做。

But it's not from a sense of like, they're doing this, so I'm going to do this.

甚至也不是因为这看起来是战略上最优的局部选择。

It's not even from a sense of like, this is a strategically optimal local thing.

而是因为我在致力于一项宏大的使命，我会所有决策都服务于这项宏大的使命。

It is like, I am working on this gigantic mission, and I will make every decision possible through the service of this gigantic mission.

这就是为什么他在战略上如此与众不同。

And that's why he's so strategically differentiated.

就像回到这个悖论：当你追求的目标是别人根本连想都不想的目标时，你所受到的激励就来自一套完全不同的体系。

Like, it is back to the paradox of like when you're doing something that shooting for a target that nobody else is even shooting for, you are incentivized by these completely different set of systems.

正是这些激励驱动了所有这些决策。

And those incentives drive all of these decisions.

这就是为什么特斯拉和SpaceX都具有相同的特质——它们都在做别人从不做的事情。

That's why Tesla and SpaceX both have the same dynamic, where they're doing things that nobody else is doing.

当你谈到这一点时，我想到两件事。

There's two things that come to mind when you're talking about that.

第一，我认为人们应该大力投资并练习提炼思想和以简单方式传达复杂内容的能力。

One, I want to talk about that I think people should heavily invest and maybe practice on their ability to, like, distill ideas and communicate, like, something complex in a simple way.

我认为埃隆可能是世界上在这方面最出色的人之一。

I think Elon's, you know, maybe arguably one of the best people in the world at doing that.

第二点，这是贯穿一切的主题，尤其是在SpaceX中最为明显，那就是将成本控制视为一种痴迷，并不是事后可以简单添加到组织中的东西。

The second part, which is a theme throughout everything, especially, and I think it's most notable in SpaceX, is like cost making cost control an obsession is not something you like graft onto the organization after.

在组织成立之前，这种理念就已经存在了。

It was there before the organization even started.

是的。

Yes.

当他产生这样一个想法时：为什么这些俄罗斯洲际弹道导弹这么贵？

When he has this idea where he's just like, Why the hell are these Russian ICBMs so expensive?

他开始运用第一性原理思考，这个故事大家都知道，很有趣。

And he starts doing the first principle thinking, which is, you know, everybody knows this story, which is interesting.

但他用另一种方式描述，我觉得这更令人难忘，他说：为什么俄罗斯会是唯一能制造出低成本火箭的国家？

But then him describing it in another way, which I found was way more memorable, he's like, Why would Russia be the only way that can make a cost effective rocket?

这在你的书里也有提到。

And it's in your book.

他接着问：我们开俄罗斯的车吗？

And he goes, Do we drive Russian cars?

我们用俄罗斯的家电吗？

Do we use Russian appliances?

根本不可能是这个国家率先掌握了制造最经济火箭的技术。

Like, there's just no way that this country figured out how to make the most cost effective launch vehicle.

他说：美国是一个相当有竞争力的地方。

He goes, America's a pretty competitive place.

我们本该拥有最经济的火箭发射系统，而事实上，他现在已经彻底主导了这一领域。

We should have the most cost effective launch vehicle, which obviously now he's like absolutely dominated in that.

但同样，用这种方式表达，你可以说：好吧，从第一性原理出发，拆解了所有部件，然后查了商品交易所的价格。

But again, it's like speaking in that manner, like you could say, Okay, well, broke down, you know, from first principles thinking, all the parts and looked up what it was on the commodities exchange.

或者你可以说，我们难道开的是俄罗斯汽车吗？

Or you could say, aren't we driving Russian cars?

你家的洗碗机是俄罗斯产的吗？

Is your dishwasher at home?

你的冰箱是俄罗斯的吗？

Is your fridge Russian?

不是。

No.

对。

Yeah.

美国制造的东西很棒，而且我们做得既出色又便宜。

It's like America makes great shit, and we make it great and cheap.

为什么我们不能在我想进入的这个领域也这么做呢？

Why can't we do the same thing for this domain that I want to operate in?

答案是，过去几十年里，这一直是一个由政府合同驱动的寡头垄断市场。

The answer is because it had been a government contract driven oligopoly for the last, like, couple decades, basically.

NASA在开创性方面非常出色，但并没有真正专注于这个项目。

NASA was amazing in pioneering, hadn't been hadn't, like, really applied themselves to this to this project.

我认为人们忘记了一点，那就是SpaceX最初的愿景并不是成立一家发射公司。

I think that is a thing that people forget is, like, the original vision for SpaceX was not start a launch company.

埃隆最初的目标是增加NASA的预算。

Elon's original goal was to increase NASA's budget.

他愿意花一亿美元用于一项慈善使命，即在火星上展示一个温室。

And he was willing to just burn a $100,000,000 on a philanthropic mission to display a greenhouse on Mars.

用现有的火箭技术把一个温室发射到火星，让这张照片登上每家报纸的头版。

So like shooting a greenhouse to Mars with existing rocket technology, get this photo on the front page of every paper.

在红色星球上有一株小小的植物，这是生命首次转移到另一个星球。

There was like a little plant on the red planet, the first time life had transferred to another planet.

他认为，这可能会增加NASA的预算，并引发一波对太空探索的兴趣和热潮。

And he's like, that will probably increase NASA's budget and like spur this wave of exploration and interest in space.

当他试图实现这一目标时，发现太空发射市场极其昂贵、缺乏创新，而且多年没有进步。

And when he tried to do that, he discovered that the space launch market was so expensive and uninnovative and hadn't been moving forward.