2022年的第11本书。第3本英文书。
荡气回肠的一本书。
读过《下一站火星：马斯克、贝佐斯和太空争夺战》，用的是人物视角，像企业家传记一样去描写商业竞争；读过《像火箭科学家一样思考》，用的是思维模型视角，从工程学的角度去拆解火箭设计和制造的第一性原理。然而这本书用的则是创业者视角，从当事人的角度去还原早期创业过程当中的艰难挑战。全书描写了从2002年SpaceX创立，到2008年猎鹰1号第四次发射终于成功的故事。
读完之后的感想如下：
首先，创业维艰。即便是强如Elon Musk，在进入一个陌生领域的时候，依然要交大量的学费：
- 2006年3月，猎鹰1号第1次发射，失败，原因是热带海岛环境（至于SpaceX如何置好好的美国本土发射场不用而必须要跑到太平洋上鸟不拉屎的夸贾林环礁岛才能发射，就又是另一个辛酸的故事了）的腐蚀导致燃料泵入口压力传感器上的一个价值5美元的铝制B型螺母开裂失效；
- 2007年3月，猎鹰1号第2次发射，失败，火箭一二级分离成功后没有进入轨道，原因是发射几分钟后水箱中的液氧晃动（变成了“泔水”）引起了致命的振荡。这个问题在模拟发射中出现过，只不过在评估风险优先级的时候被排在了第11位，而没有进入Top10风险列表。自此之后，SpaceX的所有list都要列到第11位；
- 2008年8月，猎鹰1号第3次发射，失败，在分离的时候一二级火箭发生了轻微碰撞，原因是一级火箭的引擎熄火之后，残余燃料与舱内的少量氧气相结合产生了非常轻微的却是灾难性的推力，而从引擎熄火到极间分离之间的间隔过短，导致本应开始下落的一级火箭继续前进了一点点，撞到了二级火箭的尾部。这个问题在地面模拟测试的时候由于气压较大并没有被发现，直到大气层边缘接近真空状态下才格外严重。解决方法仅仅是在软件中修改一个参数，将主引擎关闭和级间分离的时间增加4秒；
- 2008年9月，第3次发射6周之后，猎鹰1号第4次发射，几乎失败，原因是为了赶时间，SpaceX在发射前1周租用空军的大型运输机将火箭运往发射场，但没料到由于气压短时间内的剧烈变化导致火箭结构出现破损（险些爆炸），结果团队在不可能的时间段和维修条件下，在最后1周的时间内在海岛上修好并且测试了火箭，最终发射成功。
Elon Musk最初只准备了3次发射所需要的钱（其实也不够，后来拿了NASA的钱才勉强撑下去），结果第3次发射失败之后，整个团队都陷入巨大的绝望当中。那是2008年的8月，是Elon Musk人生的至暗时刻。当时的Musk已经将他的个人财富全部投入到了SpaceX和Tesla里面，两家公司同时陷入绝境，然而他几乎没有任何成果可以用来展示，也就融不到钱（SpaceX只有一个马来西亚客户）。那年夏天，他只剩下大约3000万美元的现金。朋友们警告他说不能同时支持这两家公司，而他则为如何做选择苦恼不已。“这就像有两个孩子，”Musk说，“我不忍心让任何一家公司死去。” 除了钱，他的个人生活也在崩溃 – 他需要应对无休止的负面头条新闻（例如”特斯拉死亡观察”网站上线），他和第一任妻子贾丝廷在那个夏天离婚了（他们自2000年以来一直在一起，贾丝廷生下了六个孩子，他们的长子内华达在十周大时死于婴儿猝死综合症，他们共同经历了那场可怕的悲痛），没有房子。他的新女友，一位名叫塔卢拉.莱利的英国女演员，说Musk看起来像 “死亡本身”，并描述了他如何尖叫着从噩梦中醒来，浑身疼痛。她担心他在这种压力下会崩溃，或者心脏病发作而死。
这可能就是所谓的置死地而后生。在猎鹰火箭第3次发射失败之后，Musk反而空前平静了下来：So in this dark hour, Musk chose not to play the blame game. Certainly, he could dish out brutally honest feedback, crushing feelings without regard. Instead he rallied the team with an inspiring speech. As bad as Flight Three had gone, he wanted to give his people one final swing. Outside that room, in the factory, they had the parts for a final Falcon 1 rocket. Build it, he said. And then fly it.
读到这里，所有的读者都会以为反转的时刻到了，幸福的大结局即将到来，而恰恰最富有戏剧性的，是第4次发射前的意外事故，那几乎毁掉了SpaceX。在火箭外壳被压坏之后，团队准备把它运回国修理，而在电话会议上积攒许久的压力终于爆发了，“You need to stop talking, and shut up, and listen to what I’m about to tell you,” Thompson said. “You’re not bringing that fucking rocket back. You’re going to strip that fucking thing like a Chevy. And that rocket better be fucking disassembled by the time Buzza and I get there Monday morning.”结果便是奇迹发生的时刻，在不成功便成仁的巨大压力下，团队在短短一周内完成了不可能的任务。“Of all the crazy things that we did over the years, and of all the amazing accomplishments in a short period of time, that one really stands out,” Chinnery said. “I can’t believe we disassembled an entire stage and reassembled an entire stage in the course of a week. I don’t think I could have imagined that.”They had broken virtually every rule in aerospace to pull the first stage together, but because of these heroics on Omelek, SpaceX still had one last shot at survival. They pushed on hard throughout September, working late nights under starry skies, breaking only for grilled steaks or Turkish goulash. After the pressure tests, they bolted the second stage on to the first stage. Then the launch team rolled the entire rocket—the very last Falcon 1 hardware they had to hand—out to the launchpad. By the last week of the month they were as ready as they were ever going to be.
It was fly or die. SpaceX就这样跌跌撞撞地迈向了成功。
其次，SpaceX为什么能够成功？我认为最重要的是人才。从SpaceX成立的第一天开始，Musk就把最多的精力放在了招聘上。SpaceX的前3000名员工都是他亲自面试并且拍板录用的 -Musk made hiring a priority. He personally met with every single person the company hired through the first three thousand employees. It required late nights and weekends, but he felt it important to get the right people for his company.
在早期的时候，作为行业里面的无名之辈，尤其是在之前没有任何私人商业航天公司的成功先例时，招聘是异常困难的。Elon为了招人，削尖了脑袋去参加航天圈子内各种各样的聚会，在会议上阐述自己的“让人类成为多行星物种”计划，然后被各种嘲笑。然而他要寻找的，恰恰是在嘲笑之外的few believers，他宁愿一开始就把巨大的困难和挑战亮出来，然后去挑选在这些困难挑战之下依旧可以survive下来的乐观主义者。他向5个候选人发出了offer，结果只有2个人半信半疑地接了offer，并且为了让这2位联合创始人相信公司不会在倒逼之后卷钱跑路，他还专门把他们未来2年的薪水单独存在一个第三方账户里。
Elon认为自己有三个最厉害的地方，首先是一个“超级厉害”的工程师，I’m generally supergood at engineering, personally. Most of the design decisions are mine, good or bad. 其次，是一个“非常厉害”的效率优化专家，I am very good at optimizing the engineering efficiency of a team，第三，是一个“非常厉害”的优秀工程师鉴别器，my ability to tell if someone is a good engineer or not is very good… There are a ton of phonies out there, and not many who are the real deal. I can usually tell within fifteen minutes, and I can for sure tell within a few days of working with them. （有趣的是他专门提到”Bezos is not great at engineering, to be frank”，恐怕这也是为什么蓝色起源落后于SpaceX的一个根本原因。）
Musk的面试风格十分Tough。我理解他会主要测试三个方面：技术能力、创业决心、在压力之下解决问题的能力。除了一些比较困难的技术问题之外（我之前曾经在某篇文章里看到，很多时候Elon Musk的面试过程是一个迅速的学习过程，面试者在面试结束后会抱怨自己一辈子积累的专业知识在15分钟内就被Elon给“榨干了”），他还会经常临时抛出一些出其不意的难题出来，比如“你为什么要染头发”之类的，用来测试面试者的临场反应。The engineers sitting in those seats around the conference table had to possess a certain amount of mania, too. First they had to accept Musk’s ambitious, if not all-but-impossible vision. But it takes a rarer breed still who can sprint through thickets of technical problems as someone urges them on, faster and faster. One of Musk’s most valuable skills was his ability to determine whether someone would fit this mold. His people had to be brilliant. They had to be hardworking. And there could be no nonsense.
一旦锁定了人才，Elon就会不惜一切代价把ta搞到手。在早期的时候，SpaceX的期权给的十分慷慨 – Although Musk retained a majority of shares, early hires received large chunks of stock. 不仅如此，书中还提到，为了争取某个因为妻子在Google异地工作而犹豫的候选人，他火急火燎地给Larry Page打电话，逼着Larry Page把候选人的妻子调到Google的洛杉矶办公室（SpaceX在洛杉矶）。密歇根大学航天工程系的某个教授写了一篇文章，感叹自己历届毕业生中的Top10，有一半都去了SpaceX，被Elon看到了，于是他立刻精心准备了一次招待会，邀请教授来SpaceX参观了一大圈，各种阿谀奉承之后，终于问出了最重要的问题：“剩下五个没去SpaceX的学生是谁？” 教授登时心中雪亮，“The meeting was not about me. He wanted to recruit them. He wanted the other five.”
在关键岗位上，Elon要的是最好的、经验最为丰富的人才。而在一般工程师的岗位上，他更喜欢招刚刚毕业的毕业生，因为他们年轻、简单、充满活力，可以为SpaceX付出一切。Most had no significant others pulling on their time, asking when they’d be home for dinner. They lived in apartments, not houses with lawns to mow. They had no children to look after. So they worked long, hard hours as Musk squeezed everything out of them that they had to give. And most were more than willing to give SpaceX the best years of their lives. Musk was a siren, calling brilliant young minds to SpaceX with an irresistible song. He offered an intoxicating brew of vision, charisma, audacious goals, resources, and free lattes and Cokes. When they needed something, he wrote the check. In meetings, he helped solve their most challenging technical problems. When the hour was late, he could often be found right there, beside them, working away. And when they needed a kick in the ass, he deployed his stare, or a few sharp words. 书中一位员工总结得十分到位，“We sometimes joke that SpaceX is like dog years. You get like seven years in one. And it’s true.”
和硅谷著名的车库geek风不同，SpaceX的员工大都不是码农，而是工程师，但又不是NASA里那种文质彬彬、充满古典气质的工程师，而是浑身摇滚范儿的重金属乐队。他们的Slogan是“Outsweat, Outdrink, Outlaunch”，玩儿命喝酒，玩儿命出汗，玩儿命发射。这实在是太酷了。
第三，人才带来效率。对SpaceX来说，效率就是一切。在航天业有句老话，“Faster, Better, Cheaper”是个不可能三角，你最多只能获得其中的两个。Musk不信这个邪，他全都要。通过极致的效率优化（他自我评价的另一个优点就是效率优化），他成功了。SpaceX的高效体现在以下一些地方：
- 决策效率。由于Elon自己是一个“超级厉害”的工程师，实际上他同时扮演了CEO、CTO和CFO的角色，因此在重大决策（尤其是采购决策）上可以做到非常快。This decisive style carried over into meetings back at the office in El Segundo. Musk would convene his different teams in a small conference room, be it his engineers working on propulsion, or structures, or avionics, and run down the major issues. If an engineer faced an intractable problem, Musk wanted a chance to solve it. He would suggest ideas and give his teams a day or two to troubleshoot, then report back to him. In the interim, if they needed guidance, they were told to email Musk directly, day or night. He typically responded within minutes. Over the course of a single meeting Musk could be, at turns, hilarious, deadly serious, penetrating, harsh, reflective, and a stickler for the finest details of rocket science. But most of all, he channeled a preternatural force to move things forward… When Musk decided that Spincraft could make good tanks for a fair price, that was it. No committees. No reports. Just, done.“I make the spending decisions and the engineering decisions in one head,” Elon said. “Normally those are at least two people. There’s some engineering guy who’s trying to convince a finance guy that this money should be spent. But the finance guy doesn’t understand engineering, so he can’t tell if this is a good way to spend money or not. Whereas I’m making the engineering decisions and spending decisions. So I know, already, that my brain trusts itself.”
- 自给自足。传统上，航天是一项复杂的系统工程，非常依赖上游的供应商。然而作为一个名不见经传的小玩家，SpaceX在供应商名单上的优先级是非常低的，无论是采购还是后续维修保养的成本都很高。“True, a product may already exist, but is it optimized for your solution? Is it from a good supplier? And what about their tier two or tier three suppliers? And if you need more of them faster, will they meet your needs? If you want to change something, are they going to be willing to change it? And if you improve that product, will they then sell it to your competitors?” 因此，SpaceX采用了一种自给自足的方式，Musk干脆买下了一个“臭味相投”的工厂，用来专门满足SpaceX的一切零件定制生产需求。By bringing Reagan in house, Musk essentially cut much of his manufacturing costs in half. Now he could buy a chunk of aluminum and have people in his building work it as if it were clay, producing a part on demand without the markup and delay of sending it to an outside machining shop. And the lines of communication between SpaceX’s engineers and the manufacturing crew were wide open。自给自足的方式一开始是非常痛苦的，尤其是在发动机等关键部件上，相比直接采购，自研会困难得多，是“重复发明轮子”的笨办法，但也正是因为自研让SpaceX积累了第一手的经验和关键技术，最终构建了SpaceX的核心竞争力和和护城河。When Barber-Nichols finally delivered the redesigned turbopump to SpaceX in 2003, it still had major problems. This forced Mueller and his small team to begin a crash course in turbopump technology. “The bad news is that we had to change everything,” Mueller said. “The good news is that I learned everything that can go wrong with turbopumps, and really how to fix them.” Because the pressurization of rocket fuel allows an engine to squeeze out a maximum amount of thrust, good turbopumps are essential. This would become one secret to SpaceX’s eventual dominance of the global launch market. Mueller said the original pump from Barber-Nichols weighed 150 pounds, with an output of about 3,000 horsepower. Over the next fifteen years, SpaceX engineers continued to iterate, changing the design and upgrading its parts. The turbopump in the modern-day Falcon 9 rocket’s Merlin engine still weighs 150 pounds, but produces 12,000 horsepower；
- get shit done，快速迭代。在当时的美国航天业，所有的工程项目都是线性进行的，目标被拆解成各个子目标，每个子目标要经过设计、论证、研发、试产、模块测试等一系列环节，在100%可靠的情况下才会向下一个环节推进。整个过程非常耗时耗力。而为了能够更快，SpaceX并不在前期论证上花太多的时间，而是将生产和测试的步骤大幅前置，在实践中快速发现问题，快速迭代。总之就是“少废话，干了再说”。The iterative approach begins with a goal and almost immediately leaps into concept designs, bench tests, and prototypes. The mantra with this approach is build and test early, find failures, and adapt. This is what SpaceX engineers and technicians did on the factory floor in El Segundo, and it allowed them to capture basic flaws with early prototypes, fix their designs, and build successively more “finished” iterations. 不光是在技术研发上，SpaceX的整个公司都是鲜明的“get shit done”的实用主义作风，比如销售总裁Shotwell入职后，连夜赶了一份精美的PPT，准备给Musk讲，没想到Musk扫了一眼就走了，“I was like, oh, OK, this is refreshing. I don’t have to write up a damn plan,” Shotwell recalled. Here was her first real taste of Musk’s management style. Don’t talk about doing things, just do things；
- 容忍失败。在SpaceX的发展史上，失败是常事，但Elon很少去指责员工。每次发射失败的时候，他都会在第一时间出现在全体员工面前，温言鼓励（但私下里对高管们就要严厉得多）。他身先士卒，很多时候会提出自己的看似荒谬的解决方案（“比如用降落伞来回收一级火箭”），并且不介意被证明失败。Musk had been wrong. But the filthy and exhausted engineers and technicians working with him all night did not begrudge Musk for keeping them at a task that proved fruitless. Rather, his willingness to jump into the fray, and get his hands dirty by their sides, won him admiration as a leader…Failure was an option at SpaceX, partly because the boss often asked the impossible of his team. In meetings, Musk might ask his engineers to do something that, on the face of it, seemed absurd. When they protested that it was impossible, Musk would respond with a question designed to open their minds to the problem, and potential solutions. He would ask, “What would it take?”If Musk asked Kassouf to jump a fifty-foot fence, he did not want to hear it was impossible. He wanted Kassouf to ask for a pogo stick with a certain kind of a spring on it, or maybe a jetpack, and get on with it. Musk pushed his engineers to try new approaches to difficult problems. If they had good ideas, he would back them with resources；
- 第一性原理的思考方式。The last thing Musk ever wanted to hear from an employee was “But that’s how it’s always been done.”During one early meeting at El Segundo, some former Boeing and Lockheed workers began bantering back and forth about their old companies, and the merits of how things had been done.Musk raised his voice to end the discussion. “You work at SpaceX now,” he sternly reminded them. “You bring that up one more time, and we’re going to have serious problems.”The message was clear. Wherever they had come from, whatever they had learned at those places, they were now part of the SpaceX team；
- 因地制宜，便宜行事。为了提高效率，同时还要省钱，SpaceX的团队做了不少出格的事儿，比如为了把火箭在发射场和机库间传送，他们放弃了传统的大型运输车的方案，而是用了“石器时代”的解决方案 – 自己造了个带轮子的极其丑陋（但极其实用）的“摇篮”，把火箭放在上面推来推去；在某次发现猎鹰火箭的天线进水了，他们没有用传统方法 – 把天线拆下来进行干燥处理，而是从浴室里找来吹风机直接吹干；发现海岛上的液氧产能不足时，他们在电商网站上发现了一个看起来很不靠谱的液氧制造机，但还是决定买下来试试，Elon居然还同意了。结果这台巨大的机器终于被组装起来后，足足有半个集装箱那么大，“它看起来像是从一个前苏联的疯狂科学家的实验室里出来的，阀门、仪表…所有的东西都在旋转和呼呼作响。”工程师和技术人员花了几天时间给机器上油，给它上油，给它上润滑油，最后才把它打开。它一边“打嗝”一边发出了噪音，结果当好奇的工程师们在大约45分钟后回来时，机器居然开始工作了。在没有故障的情况下，它每天生产大约300加仑，并补充了公司的L液氧供应…”“That’s the thing about Elon, he was willing to spend money to try things. And that’s so different. Go to Boeing, and you spend money to try and figure out what your liabilities are going to be before you try anything. But Elon is like, sure, try it. If it doesn’t work we can either sell it back, or it goes into our lessons-learned pile.”
最后，SpaceX的成功背后，是马斯克身上伟大的企业家精神，尤其是以下几点让我印象深刻：
- 使命感。Musk的Mars Oasis梦想是真正驱动猎鹰火箭上天的发动机。他不明白为什么NASA在把宇航员送上月球之后，美国的航天事业似乎就停滞了。在遥远的2001年，在他翻遍了NASA的网站而一无所获时，他就下定决心要飞向火星。而在之后的二十年，这种使命感与日俱增；
- 长远的规划。在Musk的脑海里，始终只有一个目标。Decisions in his world come down to a simple calculus: Will this get humans to Mars sooner, or not? Little else matters in his mind. 然而为了实现这个目标，他需要保持任务的多线程推进。让我印象深刻的是，在猎鹰1号发射倒计时30分钟的时候，Musk居然在控制台里揪着发射总指挥，一本正经地讨论猎鹰5号（下一艘火箭）的材料订购进度问题，全然不顾马上就要发射的火箭，让所有人目瞪口呆。In truth, this was just Musk being Musk, multitasking to the nth degree. Even in the middle of a critical countdown, he had the ability to simultaneously think about the company’s needs six months or a year into the future. The last thing on Thompson’s mind were shipping dates and aluminum costs. He had a rocket to launch. The company’s very first rocket, in fact. Many of the things they were doing that day were new, and uncertain. But Musk’s gaze looked far beyond the day’s launch…While trying to get Falcon 1 flying, he wanted specs for the Falcon 5. Then his small company took on the challenge of building both the Falcon 9 rocket and Dragon simultaneously. In the mid-2010s, as the company was well on its way to developing arguably the world’s best rocket in terms of price and performance, Musk pushed for rapid reuse, and then the Falcon Heavy, and a Starlink internet constellation as well as the Starship and Super Heavy Launch System.This incredible pressure wore down his employees, but for someone like Musk who sees only a narrow window to execute his sweeping vision, there is no other way.“Sometimes it would kind of get under your skin,” said Buzza, recalling Musk’s focus on the Falcon 5 rocket in the Kwaj control center, during the final countdown to that very first launch. “Like, I’m over here struggling to solve the Falcon 1, and you’re bugging us about a Falcon 5. But if you don’t have somebody like that pushing the train out to the future, your iteration time is too slow. It’s just too slow.”“Having Elon, it makes things a lot simpler because he is super involved, he makes those difficult decisions,” she said. “When those times came, he would step in and make those decisions like, Are we going or not? What are we doing here? And he’s always kept us focused on that vision. He never would relieve us of any little detailed duties, but he would always make sure that we would come out and look at the bigger picture. I think that was really important to kind of keep that focus.”
- 极致的专注。“I’ve never met a man so laser focused on his vision for what he wanted. He’s very intense, and he’s intimidating as hell”；
- 恩威并施的激励方式。In moments of high tension, Elon Musk often tries to break the stress with laughter. Musk has a rollicking wit. He will say something funny, realize it is funny, and iterate on the joke as a conversation proceeds. In this way, he brings those listening into the joke. As they sat outside, eating stew, Kimbal recalls his brother processing the day in his own way. Like other rocket enthusiasts, Musk experiences an extraordinary adrenaline rush during a launch. By the evening, he was crashing from that high, and beginning to reflect on what had happened, and what could be done to move forward. “He was clearly sad for what happened, but also making fun of the situation, because what else do you do?” Kimbal Musk said…Among Musk’s talents as the leader of SpaceX was finding different ways to motivate his employees. Steve Davis said Musk would often visit his desk to ask detailed questions about his computer simulations for controlling the rocket in flight. And then they would make bets on some aspect of the rocket and its avionics system. Almost invariably, Musk would win. But ahead of one systems test in 2007, Davis said Musk raised the stakes. Davis bet twenty dollars he could complete some aspect of the test by a certain date. In return, Musk bet a frozen yogurt machine that Davis could not make the deadline.“The second we had a bet like that, where there was a chance of getting a yogurt machine, there was a zero percent chance I was not getting that done,” Davis said. “And if you go to SpaceX today in Hawthorne, you will see that he honored the bet, and we have a frozen yogurt machine sitting in the middle of the cafeteria, which still gives away free yogurt. So yeah, he’s very good at motivating his people.”
- 可爱与疯狂的一面。为了向政府宣传自己的火箭，他直接大摇大摆地把火箭运到了国会的大门口（并且还是一艘“假”的火箭，很多部件只有逼真的壳儿而里面是空的）；在最后一次猎鹰1号发射之前，为了讨吉利，他专门带着孩子去迪士尼玩儿了“极速光轮”过山车，结果发射成功了，从此之后他就开始迷信，每次重大发射之前必然会带着孩子去迪士尼玩儿极速光轮…
所有的理性分析之余，创业故事最打动人的，始终都是那种九死一生的惊险与浪漫。读这本书的过程中，我无数次地联想到了《Cowboy Bebop》中的情节和画面，联想到人类在宇宙中是如何孤独地上下求索。书中写道，夕阳西下，猎鹰1号孤零零地矗立在德州荒凉的发射场里，一群浑身脏兮兮的工程师们结束了一天的工作，扔下了电钻和扳手，一边大笑着聊天，一边大口地喝着冰凉的的啤酒。微风吹过，飘来调频里Lynyrd Skynyrd, Aerosmith, Rolling Stones喧嚣狂躁的音乐。巨大的火箭，硬核摇滚乐，金黄色的啤酒泡沫，混合成一副末日废土的壮丽景色。
It was good living.
