2003年秋天,马斯克站在斯坦福的讲台上,台下是一群想创业的学生。他刚把PayPal卖了15亿美元,新公司SpaceX成立刚满一年,一枚火箭都还没发射过。
他没讲怎么赚钱,讲的是火箭。
火箭有个残酷的数字:起飞重量的2%能送进轨道,就算合格。Falcon就是这个水平。也就是说,你算错2%,剩下98%全掉进太平洋。
马斯克说,这就是火箭贵的根源。你得把每个数字都算对,再验算一遍,再在地面上测一遍,然后祈祷。
先把两件旧事讲完:Zip2和PayPal
1995年,马斯克在斯坦福读能源物理博士。他办了个延期入学,想试几个季度,不行就回来。
他去跟教授说这件事,教授回了一句:
我不觉得你会回来。(马斯克的导师)
那是他们最后一次对话。
95年的互联网,商业化前景一点都不明朗。马斯克在沙丘路(Sand Hill Road)见的风投,很多人压根没听说过互联网。听着离谱,但那会儿就这样。
他还是决定干。理由很朴素:这种事很久才来一次。
问题是没钱。所以第一家公司必须能很快产生现金流。他观察了一圈,媒体行业需要把内容搬到网上,而且他们有钱。当时互联网上还没有广告收入,帮媒体做数字化转型,是唯一看得见的付费方。
Zip2就这么来的。他们给报业写了不少软件,客户和投资人包括赫斯特(Hearst)、奈特瑞德(Knight Ridder)这些美国主流印刷出版商。1999年初,康柏(Compaq)提出收购,价格是三亿多美元现金。
马斯克对现金这个币种评价很高:
现金是我强烈推荐的一种货币。(马斯克)
PayPal的起点是另一个判断。1999年初,他环顾互联网,发现金融服务业几乎没什么创新。
他的推理是:钱是低带宽的东西。你不需要什么基础设施大升级才能处理它,它本质上就是数据库里的一条记录。纸币只占全部货币的很小一部分。这么简单的东西,理应最适合在互联网上被重构。
他们想了两个方向。一个是把所有金融需求整合到一个网站:银行、券商、保险,全在一页。另一个是邮件转账,输入一个邮箱地址或任何唯一标识,就能把钱、股票、基金从一个人转给另一个人。对方没账户,系统会自动发邮件邀请他注册。
第一个方向非常难,他们解决掉了大部分问题。第二个方向花了一天。
然后他们拿着两套功能去演示。讲第一套,说我们花了多大力气,你看银行账单、基金、保险全在一页,多方便。观众反应平平。讲第二套,说输入邮箱就能转账,观众的反应是"哇"。
方向就这么定了。
当时公司叫X.com。另一家公司叫Confinity,起点完全不同:他们做Palm Pilot上的加密,演示程序是用红外口在掌上电脑之间"发射"代币付款,配套网站叫PayPal,用来对账。结果他们发现,网站部分比掌上电脑加密有意思得多,业务重心就转过去了。2000年初,X.com收购Confinity,一年后公司改名PayPal。
PayPal是病毒式营销的教科书案例,和Hotmail一个路数。一个客户会顺带变成你的销售员,他给朋友转钱,等于把朋友拉进网络。客户越多,增长越快,像培养皿里的细菌,一条S曲线。马斯克管了PayPal头两年,第一年上线,第二年底用户到一百万。公司没有销售团队,没有营销团队,广告费是零。
2002年2月PayPal上市,是那半年唯一一家上市的互联网公司。过程不算顺利,马斯克说自己公司创下了SEC(美国证券交易委员会)要求重写招股书的纪录,正好赶上安然事件,监管把所有人都往死里审。
同年6、7月,eBay出手收购,价格约15亿美元。当时eBay股价55美元,还没拆股,折算成今天的市值大概30亿。
阿波罗之后,人类在太空原地踏步
也是在2002年上半年,马斯克开始做一件看起来毫不相干的事:研究航天。
他想弄明白一个问题。60年代,美国从零开始,连把人送上天都做不到,十年之内把人送上了月球,所有技术都是现造的。然后70年代、80年代、90年代,这条路就横向走了。到2003年,美国连把人送进近地轨道的能力都没有。
这跟其他所有技术领域都对不上。70年代初你能买到的电脑,体积能填满这间教室,算力还不如今天的手机。几乎每个技术门类都在进步,航天没有。
他一开始以为这是钱的问题。钱要靠公众支持来撬动,那就要先让公众兴奋起来。
于是他设计了一个任务,叫"火星绿洲"(Mars Oasis)。往火星表面放一个小型机器人着陆器,带种子和脱水营养凝胶,着陆后水合,让植物在火星的辐射和重力条件下生长,同时在地表维持一套生命支持系统。
成本估下来1.15到1.2亿美元,大约是NASA低成本任务的十分之一。
这个设计的逻辑是,公众对"第一次"和"最"有反应。这会是生命到达过的最远地方,也是火星上的第一批生命。
计划卡在了运载环节。美国当时最便宜的火箭是波音的Delta 2,一次5000万美元。对这个小任务来说,太贵了。
莫斯科买导弹:谈成了,但他不敢签
马斯克去了三趟莫斯科,想买一枚俄罗斯的运载火箭,本质上是翻新的洲际弹道导弹。
他自己都说,去莫斯科谈 refurbished ICBM(翻新洲际导弹),在"人生奇特经历"排行榜上排得很靠前。
生意最后谈成了。但交易牵扯的复杂因素太多,他不愿意承担那些风险,没签。
回程的飞机上,他换了个角度想问题:为什么俄罗斯能造出便宜的运载火箭?
我们不喝伏特加以外的俄罗斯产品,不开俄罗斯车,不坐俄罗斯飞机,不用俄罗斯厨房电器。美国是个竞争很充分的地方,按理说也能造出成本可控的运载火箭。
于是他组织了一次可行性研究。参与者是过去三十年里参与过所有主要运载火箭开发的工程师,2003年初的几个周六反复迭代,目标不只是降低发射成本,还包括提高发射可靠性。最后拿出了一份初步设计。
时间赶得很巧,研究收尾正好是PayPal卖给eBay的时候。交易一完成,他就搬去了洛杉矶。全球航空航天产业最密集的地方在南加州,它是那里最大的产业,比娱乐业大得多。他在帕洛阿尔托住了九年,对加州不陌生。
航天飞机的账:一年40亿美元,飞四次
讲SpaceX之前,马斯克先把当时的行业现状摊开讲了一遍。
美国政府的载人航天状态不好。剩下三架航天飞机全部停飞,首飞时间看起来要等一年。这个飞行器极其昂贵,也相当危险。
他列了几条设计上的问题。乘员舱是侧挂的,一旦爆炸,基本是瞬间死亡。固体火箭助推器一旦点火就关不掉。燃料和氧化剂预混,本身就带着根本性的危险。再入时要靠机翼和控制面维持精确攻角,攻角瞬间偏差就能让整架飞行器解体。再加上没有逃逸系统,出事就是死。
成本更难理解。航天飞机项目把所有环节加起来,一年约40亿美元。用40亿除以一年飞几次,就是单次成本。如果一年飞四次(已经很久没飞这么多了),一次就是10亿美元。
短期计划是继续建空间站,所以航天飞机还得飞,但大概只会保持最低飞行次数。
长期方案是造一个"轨道太空飞机"(Orbital Space Plane)。马斯克说"飞机"两个字要打引号,因为其中一个方案根本就是胶囊舱,叫"椭圆形太空物体"更准确。核心诉求是比航天飞机便宜一点、安全很多,尤其是要有逃逸系统,出问题能中止。
问题是它也没便宜到哪去。单次飞行估算3到4亿美元,其中2亿是付给波音的Delta 4重型一次性助推器。整个开发要50亿美元,预计9到10年完成。按历史先例,马斯克认为它既守不住50亿预算,也守不住时间表。
对比之下,俄罗斯的联盟号是当时唯一能往返空间站的工具。便宜得多,也安全得多。乘员舱在顶部,有逃逸系统,没有机翼和控制面,少了很多能出错的东西。单次成本约6000万美元,比航天飞机低一个数量级。
限制俄罗斯的是经济。一个比利时体量的经济体,很难支撑宏大的航天计划。
最有意思的变量是中国。就在那个月,中国预计会把第一位航天员送入轨道,成为史上第三个做到这件事的国家。中国在这个项目上投了很多钱和精力,而且目标不止近地轨道,规划里有空间站、火星基地,最终是载人登火星。
马斯克说,如果说有什么能刺激美国的载人航天,大概就是中国在太空的野心,以及美国想至少不落后的那点心思。
四个人在造火箭
马斯克认为,美国正在发生的事里,最终可能超过上面所有政府项目的,是商业航天。
他给了一个类比。互联网早期,DARPA(美国国防高级研究计划局)出钱开发了大量基础技术,后来商业部门进来,才有了那种爆发式加速。NASA可能扮演了同样的角色:先花钱把基础技术做出来,等自由企业进场,曲线就陡了。
当时有几家认真的运载火箭公司在做这件事。
第一家是伯特·鲁坦(Burt Rutan)的Scaled Composites。鲁坦是世界上最顶尖的飞机设计师之一,他们造的亚轨道飞行器已经在莫哈韦试飞。母舰叫White Knight,挂在腹部的是SpaceShipOne,属于X Prize级别,据说由保罗·艾伦(Paul Allen)出资。马斯克专门强调,进入这个领域的资本相当可观。
这套架构的缺点是不可扩展,做亚轨道很合适,要做入轨飞行器,架构得重来。
第二家是约翰·卡马克(John Carmack)。写《雷神之锤》(Quake)和《毁灭战士》(Doom)的那位,马斯克评价他是世界上最好的软件工程师之一,也是他认识的最好的工程师之一。卡马克在做垂直起降飞行器,没有航天工程背景,基本靠自己加三个哥们。马斯克说他的进展相当惊人,做出来的东西能飞。但要入轨,质量效率和发动机效率都得大幅提升,大概率得做成两级。
第三家是杰夫·贝索斯(Jeff Bezos)。他是太空的铁杆支持者,高中毕业典礼上的致辞讲的就是人类必须扩展到其他星球。据说他打算未来20年在太空探索上投10亿美元量级。
第四家是SpaceX。
马斯克说,SpaceX花的钱比前面三家加起来还多,某些方面可能多一个数量级,因为做的事情难一个数量级。他们要造的是入轨运载火箭,两级,发动机效率极高,质量效率极高。
商业策略是先做卫星发射。这个市场需求是确定存在的:把中小型卫星送进轨道。以此建立收入基础,再进入载人运输。公司的长期目标明确是载人。最终的升级路径是土星5号的继承者,一枚超重型运载火箭,用来建月球基地或者执行火星任务。他把这个叫"圣杯"级别的目标。
技术细节上,主发动机叫梅林(Merlin),海平面推力约75,000磅。上面级发动机在真空中的推力约7,500磅。
首飞安排在2004年3月左右,从范登堡空军基地(Vandenberg Air Force Base)的SpaceX发射场起飞,载荷是一颗海军通信卫星。马斯克特意点出这件事的分量:运载火箭公司通常没法在首飞就拿到付费客户,他们拿到了。
他还说了一句,Falcon是史上开发速度最快的运载火箭,包括战时。
火箭为什么贵:2%的生死线
有人问,为什么把东西送上太空这么贵。
马斯克的回答从物理开始。入轨需要的能量和速度太大,跟汽车、飞机完全不是一个量级,几乎没有余量可以浪费。一枚运载火箭能把起飞质量的2%送进轨道,Falcon也是这样。
这意味着,如果你算错2%,你什么都送不上去,剩下的部分掉进太平洋。
所有计算必须正确。算错一个数,火箭就炸。而要让所有答案都正确,还要复核、还要在地面上尽可能多地测试,这些全都烧钱。
第二个原因是发射频率低。如果一年能飞几千次,成本会低得多。但这里有鸡生蛋的问题:得先便宜,才能有几千次飞行。
马斯克的结论是,火箭本来就该比现在便宜得多,现在的造法和运营方式效率太低了。
Falcon的售价定在每次约600万美元。最接近的竞争对手是轨道科学公司(Orbital Sciences)的飞马座(Pegasus),每次约2500万美元,能力还不如Falcon。这算是进入太空成本上的一次实质性突破。
便宜不是靠银弹,是靠几百个小改动
有人追问,价格是怎么降下来的。
马斯克的回答是,没有银弹。没有任何一项改动单独贡献了大部分成本节省,是几百个小创新叠加起来的结果。推进系统、结构、航电、发射操作,每一块都改,再加上一个开销极低的组织。
他先讲了组织。30人的公司,管理开销比洛克希德或波音低一个数量级。就算其他什么都不改,只造同样的火箭,光这一项就已经便宜很多。
然后讲原则:每个决定都要考虑简单性。简单同时提升可靠性和降低成本。零件少,能坏的少,能卡住的也少。
机身是个例子。半硬壳式结构,蒙皮厚度可变,共用一个隔板。马斯克说自己得画张图才能讲清楚,但结果是便宜、质量效率高、好测试、可靠。
航电是另一个例子。他们在火箭内部用以太网通信。这在其他行业听起来不算什么,在运载火箭里算是创新。其他火箭用的是贯穿全箭的串行电缆,铜线束粗得像手臂,又重又贵。这类改动一个个加起来,差距就出来了。
卖火箭和卖支付,是两个世界
有人问客户群的差别。
马斯克说差别巨大。PayPal是消费品,SpaceX卖的是火箭,想买火箭的人非常少。
如果这里有谁有600万美元又想要一枚火箭,我很乐意卖给他。(马斯克)
销售过程高度个人化。买一次发射要做的思考,比注册一个PayPal账户多得多,后者几乎不花你什么成本。
至于病毒式营销:
火箭大概不会有什么病毒式传播。我希望有,但我不指望。(马斯克)
两家公司共用的那套做法
被问到Zip2和PayPal有什么共同点时,马斯克列了几条。
两家公司的技术核心都是软件。Zip2服务媒体行业,PayPal服务金融行业,内核都是软件和互联网。
两家都在帕洛阿尔托,他住的地方。
两家都是小团队加高人才。PayPal人员最多的时候大约30个工程师,马斯克说这套系统比美联储的清算系统更复杂,而且他很确定这一点,因为美联储那套系统很烂。
组织是典型的硅谷创业公司:层级很平,工位大小差不多,谁都能跟谁说话。哲学是最好的想法胜出,而不是提出想法的人因为职位高就赢。马斯克补了一句,虽然有时候他觉得应该反过来。
所有人都持股。
决策上有个习惯:如果两条路摆在那,没有一条明显更好,那就别花大量时间论证哪条好2%,直接选一条做。有时候会选错,走了次优路径。但选一条路走下去,通常比无休止地摇摆强。
知识产权文书、法律流程,他们不太上心。精力全放在把产品做到最好。两家公司都是产品导向,对客户体验近乎偏执。马斯克认为这比养一支庞大销售团队、想营销花招、搞十二步流程有效得多。
说到病毒式营销,他给了个定义:一个客户替你把另一个客户带来,而你什么都不用做。Friendster、Hotmail、PayPal、eBay都是这样。
这种模式里,产品重要到极点。因为你要把东西推荐给朋友,你必须真的喜欢那个体验,否则你不想烧掉这段关系。
还有一条关于并行的经验。PayPal要对接一堆后台:信用卡系统、美联储的电子资金转账系统、各种欺诈数据库。每一样都要谈,都要接。前后花了大约一年。
他的总结是,尽量不要把依赖串行化。能并行的事情全部并行。很多事有孕育期,你没法加速,那至少让它们同时孕育。人倾向于把事情排成队,这是拖慢进度的常见原因。
关于专利,PayPal申请了几项,但从来没有真正起过作用。马斯克认为专利主要是防御工具,不是进攻武器,主动起诉别人非常难。制药这类行业专利极其重要,软件生命周期太快,你拿到专利的时候它可能已经过时了。
硅谷是自由意志主义天堂,航天不是
被问到进入航天业的障碍,马斯克先讲监管。
这套监管结构非常复杂。你造一枚入轨运载火箭,它和洲际弹道导弹没有本质区别,所以监管多是合理的,毕竟没人想看到发射的东西掉在洛杉矶。
真正难的是环保审批,比他们预想的难得多。
马斯克在这里说了一段关于硅谷的观察。在硅谷住久了,你会意识到自己活在一个自由意志主义的天堂,几乎没有监管。而出了这个环境,最让人抓狂的是监管常常不讲道理。你面对的人在执行一套规则,至于规则本身合不合理,不在他的考虑范围。你可以试着说服他规则不合理,他不听。
还有一类限制来自安全。SpaceX只能雇有绿卡或是美国公民的人。马斯克的说法很直白:如果没法把你关进监狱,他们就不让你碰火箭相关的工作。跟任何外国公民谈话,都需要国务院的技术转让协议。他们的第二次发射是非美国政府客户,光拿到国务院批准、可以开始谈合同,就花了六个月。
公司没有律师。监管文件技术性极强,马斯克说这更像给FAA做飞机认证,只不过认证对象是火箭。他开玩笑说希望能把这些活外包给律师,但律师接不了。
太空采矿和太空太阳能,账差三个数量级
有人问太空里哪些商业模式靠谱。
马斯克直接点了两个他认为不靠谱的:太空采矿和太空太阳能发电。
算法很简单。不管是把太空太阳能发的电(光子)传回地球,还是把原材料运回地球,成本都压不下来。经济账差着大约三个数量级,怎么算都闭合不了。
他认为真正大的机会在别处:在月球或火星上建立基地,尤其是建一个能自给自足的文明。那是万亿美元量级的生意,因为到那一步,本质上就是行星际商业在运转。
但他强调,不会是太空太阳能,也不会是太空采矿。
什么人适合创业
最后有人问创业者该有什么特质。
马斯克的回答是,成功的创业者什么形状、什么口味都有,没有统一模板。他只能说对自己重要的几条。
第一条是对产品质量的偏执。在这个语境里,强迫症是优点。
第二条是真心喜欢你做的事。就算你是最强的那个,失败的概率依然存在,所以得真的喜欢。不喜欢的话,人生太短。而且如果你喜欢,你不工作的时候也会想它,脑子会自动往那儿跑。不喜欢,就很难做成。
关于沃顿的商学位,他说商学院教了很多术语,把一些你本来得靠实践摸索的概念提前介绍给你,这能加速学习过程。但他也见过高中都没毕业的成功创业者,也见过博士。核心不是学历,是致力于学会你需要会的东西,在学校学或在实践中学都行。
他顺带讲了Zip2最惨的那段。
1995年夏天他写软件,做了个能在网上查地图和路线的工具,还有个能在线编辑内容的系统,相当于很先进的博客工具。然后开始联系小报纸和媒体公司。一半的人反应是"什么是互联网",在硅谷也是这样。偶尔有人听懂了,能给一点钱。
团队六个人:他自己、被他从加拿大劝来的弟弟、他妈妈的一个朋友,外加三个在报纸上登广告招来的提成制销售。
他没有钱,是负资产,背着巨额学生贷款。租不起住的地方,也租不起办公室。他租了办公室,因为办公室比住处便宜。然后睡在办公室的沙发上,去Page Mill和El Camino那家YMCA洗澡。
他说那段时间是他身体最好的时候,洗澡加锻炼,一天就齐了。
楼下一层有家很小的ISP,他们在地板上打了个洞,接了一根null modem线,一个月100美元就有了网络。烧钱率低到荒谬。收入也很少,但确实比支出多一点。
所以去见风投的时候,他们是有底气的。
至于太空这条路,有人问他会不会推荐首次创业者去做。他说不会。这是高阶创业。他听过太多遍那句话:
在航天业赚一笔小钱最快的方法,是先有一大笔钱。(马斯克)
他希望这句话在自己身上不成立。但对第一次创业的人来说,太空太硬了。资本需求太高,不如从需要少量资金的事情做起。
把这场2003年的演讲读完,会发现它最有价值的地方不在于预测对了什么,而在于展示了一种思考方式:先算物理和成本,再谈愿景。二十年后回看,SpaceX的路径几乎就是这场演讲里写的剧本:先做卫星发射建立收入,再做载人,最后做超重型运载。
如果你想从这场演讲里拿走点能用的东西,可以试试这几条:
- 遇到两条没有明显优劣的路,选一条立刻做,别花几周论证哪条好2%。走错再改,比原地摇摆便宜。
- 把所有能并行的依赖全部并行。很多事有固定的孕育期,你唯一能控制的是让它们同时开始。
- 降成本先看组织开销和零件数量。30人公司的管理成本比巨头低一个数量级,简单性同时买到可靠性和低价。
- 做消费品就死磕产品体验,因为用户推荐你的时候是在拿朋友关系做担保。做企业级生意就别指望病毒式传播,老老实实一个个谈。
- 进一个重资产、强监管的行业之前,先算清楚监管成本和时间成本。环保审批和跨部门批文往往比技术本身更耗人。
- 判断一个商业模式能不能成立,先把单位经济算一遍。太空采矿和太空太阳能听着性感,账差三个数量级,就是不成立。
- 选方向的时候问自己一句:不工作的时候,脑子会不会自动想这件事?答案是不会,就换一个。
0:01I'll try to make this as interesting as possible if you like space you'll like this talk so my background in brief I'll I'll talk a little bit about zip 2 and and PayPal and then mostly about space and what we're doing in space so I originally came out of California to do energy physics at Stanford actually and ended up putting this is in ' 95 and ended up putting that on hold to start zip 2 and ZIP 2 I'll tell you a little
0:35bit about the thought process of exactly what happened there in '95 the it wasn't at all clear that the internet was going to be a big commercial thing in fact most of the Venture capitalists that I talked to hadn't even heard of the internet which sounds bizarre on San Hill Road so but but I wanted
0:56to to do something and in the I thought it would be a pretty huge thing and I thought it was one of those things that only came along once in in a very long while so I I got a deferment at Stanford and I thought I'd give it a couple quarters and if it didn't work out which I thought it probably wouldn't then I'd come back to school and actually when I talked to my professor and I told him this he said well I don't think you'll be coming back and that was the last conversation I had with him so but there weren't a lot of ways to get the only way to get involved with the internet in '95 that I could think of was to start a company because there wer a lot of companies to go and
1:37work for apart from Netscape maybe one or two others so and then I didn't have any money so I thought well we got to make something that's going to return it's going to return money very very quickly so we thought well the the the media industry would need help converting its content from from a media to electronic and they clearly had money so if we could find a way to help them move their media to the internet that would be an obvious way of of generating Revenue there was no advertising revenue on the internet at the time so and that that was really the basis of of zip 2 and we ended up building quite a bit of software for for the media industry
2:21primarily the print media industry so we had as investors and customers H Corporation night rder Mo most of the the major US print Publishers we built that up and then we we had the opportunity to sell to compact in in early 99 and basically took that offer it was for a little over $300 million in cash and that's that's a currency I highly recommend so so we had that and and
2:56but but I wanted to do something something more after zip two so post the sale in fact immediately post the sale I didn't really take any time off I was trying to think of where where the were the opportunities in this is early 99 where the opportunities remained in the internet and it seem to me that there hadn't been a lot of innovation in the financial services sector so and and when you think about it money is is low bandwidth you don't need some sort of big infrastructure Improvement to to do to do things with it it's really just an entry in the database the the paper form of money is is really only a small percentage of all the money that
3:35that's out there so it should land itself to Innovation on the internet and so we thought of a couple of different things we could do one one of the things was to combine all of somebody's Financial Services needs into one website so you could have banking brokerage insurance and all sorts of things in one place and and that was actually quite a difficult problem to solve but we we we we we solved most of the issues associated with that and then we had a little feature which took us about a day that was the ability to email money from One customer to another so you can type in an email address or actually any unique identifier and transfer funds or or conceivably stocks or
4:18mutual funds or whatever from from one account holder to another and if you try to transfer money to somebody who didn't have an account in the system it would then forward an email to them saying hey why don't you sign up and open an account so and whenever we demonstrate these two sets of features we'd say well this this was a feature that took us a lot of effort to do and look how you can see your bank statement and your your mutual funds and and insurance and all that it's all on one page and look how convenient that is and people go ho home and then we'd say and by the way we have this feature where you can enter somebody's email address and transpose funds and they go
4:56wow so so we we like okay so we focused the the company's business on on email payments and and in the early G Going the company was called x.com and then there was a there was another company called confinity which had actually also started out from a from a different area they started off with the Palm Pilot cryptography and then they had as a demo application the ability to beam token payments from one Palm Pilot to another by the impared court and then they had a a website which is called PayPal where you would reconcile the beamed payments and and what they found was that the the website portion was actually far more interesting to people
5:38than the than Palm Pilot cryptography was so they they started leading their business in that direction and then in basically early 2000 x.com acquired confinity and then about a year later we we ended up changing the company's name to PayPal and that's that's kind of how the approximate evolution of of the company and but PayPal is really a case of of of where the the whole viral marketing U it's really a perfect case example of viral marketing like hot mail was where one one one customer would essentially act as a salesperson for you for all for for bringing in other customers so they would send money to a friend and essentially recruit that friend into the
6:30network and so you had this exponential growth the more customers you had the faster it grew so you it was like bacteria in a Petri dish you know it just goes like does this S curve and in fact I I ran PayPal for about the first two years of its existence and we launched after year one and by the end of year two we had a million customers so gives you an examp you know to gives you a sense of how fast things things grow in that scenario we were and we didn't have a sales force we actually didn't have a BP of sales we didn't have a BP of marketing and we didn't spend any money on advertising in about February of last year for those some some of you probably
7:17following it in February last year PayPal went public and we were I think we were the only internet company to go public in the first part of last year it went off reasonbly ly well although I think we had more SEC rewrites than than any company I can imagine like we set a record on SEC rewrites this was right around the Enron time and when there was all sorts of corporate scandals so they put us through the ringer and then shortly thereafter in
7:48about June July we struck a deal with eBay to sell a company to eBay and pH about and half billion dollar but that was when eBay's stock price was about $55 and they hadn't split so I guess in today's dollars we about $3 billion so that worked out pretty well it happened coincidentally that in in the first part of last year I've been doing just some background research on on space and let me talk a bit about that so essentially I was trying to figure out why we had not made more progress since since Apollo in in 19 in the 60s we went from basically nothing not being able to put anyone into space to putting people on the moon and developing all the technology from
8:46scratch to to do that and yet in the the 70s and 80s and the 90s we we've kind of gone sideways and and we're currently in a situation where we can't even put a person into into low earth orbit and that that doesn't really gel with all of the other technology sectors out there the computer that you could have bought in the early '70s you know would have filled this room and had less computing power than your cell phone and so just about every sector of technology is improved why is this not improved so I started looking into that initially I thought well perhaps it's a question of funding and that funding can be garnered by by really marshalling public support forward so
9:30thought well one way to to to get the public excited about space would be to do maybe a privately funded robotic Spa Mission Tom Mars and so we we figured out a mission that would cost about1 15 to20 million which isn't a lot of money but it's about a tenth of what a lowcost NASA Mission would be and and the idea was was called Mars Oasis where we' put a small robotic Lander on the surface of Mars with seeds and dehydrated nutrient Jael there would hydrate upon landing and you'd have plants growing in Martian radiation and gravity conditions and You' also be maintaining essentially a life support system on the surface of Mars and this would be interesting to the public
10:11because they tend to respond to precedents and superlatives and this would be the furthest at life has ever traveled and the first life on Mars so pretty significant and then when I start looking at at launch vehicles the lowest cost vehicle in the US is Boeing Delta 2 which costs about 50 mli million and and that's a bit that's a bit steep what we were trying to do so I made three visits to to Moscow to Russia to look at at buying a a Russian launch and it's it's actually pretty interesting going to to Moscow to negotiate for a refurbished ICBM
10:52if you know on the range of interesting experiences that's pretty pretty far out there but but we actually did get to to a deal but there were so many complications associated with the deal that that I it didn't I wasn't comfortable with the risks associated with it so when I got back from the third trip I thought well you know why is it the Russians can build these lowcost launch Vehicles because it's not like we we drive Russian cars fly Russian planes or have Russian kitchen appliances and you know when's Last bought something Russian that wasn't vodka so I think the US is a pretty competitive Place and and we should be able to to build a a cost
11:38efficient launch vehicle so I I put together a feasibility study which consisted of of Engineers that have been involved with all the major launch vehicle developments over the last three decades and we iterated over a number of Saturdays in the beginning of last year to figure out well what would be what would be the smartest way to approach this problem of not just launch cost but also launch reliability and and we came up with a default design and and that actually sort a fortunate timing that feasibility study finished up right around the time that we agreed to sell PayPal to eBay so coincident with that sale I moved down to to La where there's actually the biggest
12:19concentration of Aerospace industry in the world it's actually the biggest industry in in Southern California much bigger than entertainment or anything else I I was living in in palala for about nine years before that anyway so just to talk little broadly about space and where things are today obviously US Government man's exploration is not in a great place we've got the the three remaining shuttles are grounded it looks like first flight might only be a year from now if that and and we've got a vehicle that is incredibly expensive and and and really quite dangerous it's you know for reasons mentioned there it's got a side mounted crew compartment
13:03so if there's an explosion it's basically instant death the you've got solid rocket boosters which once you ignite them you can turn them off and there's something fundamentally dangerous about premixing your fuel and oxidizer I think and then it's you've got wing wings and control surfaces when you when you reenter you've got to maintain a precise angle of attack even a momentary variance in that can break the whole vehicle apart so so and and and then of course it's got no Escape system so if anything does go wrong your toast and and you've got a cost that is is really pretty hard to Fathom the the the shuttle program when you add up all the pieces is about 4
13:42billion a year and so you can divide 4 billion by the number of flights and that'll tell you what the cost is and and if if there's say four flights a year which they haven't been for a while then you're talking about a billion dollars a flight the plans for the immediate future obviously we've got to continue building the space station so we're going to keep flying the shuttle but but I think it's probably going to be the minimum number of shuttle flights that we need to launch the the long-term plans I would build something called orbital space plane and I'd say plane in quotes because one of the options is a capsule so it should be called you
14:25maybe oval space thing but but the basic idea is to have something that's hopefully a little cheaper and a lot safer than the space shuttle so in particular it's going to have an escape system so if something does go wrong you can abort to safety the the downside is that it's still
14:47it while it might be a little cheaper it's still going to be pretty nonexpensive the estimated cost per flight of overal space plane is somewhere in the region of 300 to $400 million a flight and of that that amount 200 just $200 million alone is goes to Boeing for the Delta 4 heavy Expendable booster and and the it's a $5 billion
15:10development effort and expected to be completed in 9 or 10 years now typically typically things have have not been under budget and under time so it's unlikely I think given historical precedent that that it will St within $15 billion and the 20122 timeline and a bit about what's going on elsewhere in the world in Russia the this is really the soy is our only access to space station it's considerably cheaper considerably safer the so has a very good track record it's the crew is top mounted it has an escape system there are no wings or control servs just to go wrong overall it's a it's a pretty good system and the estimated costs are about $60
16:01million a flight which is an order of Mag more than order of magnitude less than than the space shuttle the thing that constrains them obviously is the weakness of the Russian economy it's very hard for them to embark on ambitious programs with an economy the size of Belgium so China is is probably the most interesting thing that's that's going on in space the this month China is expected to be only the third to it is expected to launch their first person into space that'll make them only the Third Country ever to put someone in orbit and they've put a lot of of money and effort into this program if if anything serves as a spur for human space exploration it is likely to be
16:45the China's Ambitions in space and and and hopefully a sense in America that we want at least keep up with with China and they have grand ambitions Beyond just lowth orbit they're planning on setting up a space station putting a base on Mars and eventually sending humans to Mars so what's happening in in in the US that I think might ultimately surpass all of that stuff is entrepreneurial space activities where things are led by the spirit of free enterprise and I think there's perhaps an analogy here where just as DARPA served
17:29as the initial impetus for for the internet and underwrote a lot of the costs of developing the internet in the beginning one it may be the case that NASA has is has essentially done the same thing by spending the money to develop some of the fundamental Technologies in the beginning and then once we can bring sort of the commercial free enterprise sector into it then we can see the dramatic acceleration that we did that we saw in the internet so there there are several serious launch vehicle efforts underway talk about each one there's the Bert ran of scaled composits Bert ran is is one of the world's foremost aircraft designers and he's developed
18:10a subal vehicle that they're actually flying out of Mojave this is an x- prise class vehicle there's John KAC who wrote quake and doom was probably one of the best software engineers in the world one of the best Engineers that I know period and he's developing a vertical tle from land vehicle Jeff Bezos who I understand was here last week is a huge space ad Advocate
18:37and it's my understanding that he intends to spend something in the order of a billion dollars over the next 20 years on space exploration and and then space then my company SpaceX and I I think within the next several years these entrepreneural efforts will actually be what what deres space exploration so a little bit about each one that's a picture of of the veran F it it's called White Knight is the carrier plane and then spaceship one is the thing that's held in the belly there and this project is supposedly funded by Paul Allen so there's quite a lot of capital also I should make that point quite a lot of capital that's sort of entering this
19:21entrepreneurial space sector as well the I the only drawback I think of this architecture is that it's not really scalable for for something that would get to orbit this is pretty good for suborbital but I think the architecture would need to change for an orbital vehicle and there's John KX SE you see he's a little irreverent this is from his website his his vehicle is a vertical takeoff and Landing vehicle he's made he's made really incredible progress for somebody who has no background in
19:58aerospace engineering and is also kind of doing it all himself with him and like three buddies so and and I think that will make something that works actually if you you can check out their website alad Aerospace it's pretty interesting stuff but in in order to get or to orbit this would require a substantial Improvement in the mass efficiency and the engine efficiency and probably be a two-stage vehicle Jeff Bezos who I'm sure almost everyone here is aware of he is a pretty huge fan of space in fact his High School B dictorian speech was about the necessity of humanity expanding to other planets so it's it's pretty important to him from what an
20:48that this is our effort we're spending quite a bit more than the three prior entities that I mentioned in some cases probably an order of magnitude more because what we're doing is really an order of magnitude more difficult we're building an orbital launch vehicle it's a two-stage very high efficiency engines very high mass efficiency launch vehicle and it's targeted to the satellite delivery market so our approach is really to make this a solid Sound business and so I predicated the the Strategic plan on a known Market something that I we know for a fact exists which is the need to put small to mediumsized satellites into orbit and so that that's what we're
21:33going after initially and then with that as a as kind of a revenue base we will move into the human Transportation market so long-term aims of the company are definitely human Transportation I think this the smart strategy is to First go for cargo delivery essentially satellite delivery and our eventual upgrade path is to fold the successor to Saturn 5
21:58build a super heavy lift vehicle that could you know be used for setting up a moon base or doing a m mission that would be the Holy Grail objective on the upper right there you can see a test firing of our engine and on the lower right you can see the Alo stage tank this is an engine test of our main engine which is called Merlin
22:40and that that generates about roughly 75,000 lbs of thrust at sea level this is our upper stage engine it's about it's about 7,500 lb of thrust in vacuum and this is a a accelerated version of our launch sequence our first launch will be from the SpaceX launch complex at Vandenberg Air Force Base and approximately March of next year basically early next year and we will be flying a a Navy satellite u Navy communication satellite so it's it's notable because often launch vehicle companies are not able to get somebody get a paying customer on their first flight but we've we've been able to do that this is also the Falcon development at SpaceX is the
23:40fastest launch vehicle development in history including wartime so that's actually vburg Air Force Base which is about two hours away from Santa Barbara common themes between zip 2 and PayPal well I guess both of them involved had had software at the as the heart of the technology even though zip 2 was in
24:13was servicing the media sector and obviously PayPal was servicing the financial sector the the the heart of it was really software and and internet related stuff so certainly that's a a huge commonality they were both in paloalto where I lived I think we you know took a similar approach to building both companies which was to have a small group of very talented people and and keep
24:49it small you know I think PayPal had at at its height probably 30 Engineers for for a system that's I would say is more spe ticated than the Federal Reserve clearing clearing system I'm pretty sure it is actually because the Federal Reserve clearing system sucks so what else is there it
25:12it generally you know I think the the way both zip2 and PayPal operated was was really your it was your canonical Silicon Valley startup you know pretty flat hierarchy you know everybody had a roughly similar Cube and anyone could talk to anyone we had a philosophy of best idea wins as opposed to you know the person proposing the idea winning because they are who they are even though there were times when I thought that should have been the way go and obviously you know everyone was was an equity stakeholder give us anything if if if there were two paths that you know let's say we had to choose to do one thing or the other and the one wasn't obviously better than the other
26:09then rather than spend a lot of time trying to figure out which one was slightly better we would just pick one and do it and sometimes we'd be wrong and and and we pick this sub optimal path but often it's better to pick a path and do it than to just vacillate endlessly on on a choice we we didn't worried too much about intellectual property paperwork legal stuff we we really were very focused on building the best product that we possibly could both Z two and PayPal were very product focused companies you know that's we're just incredibly obsessive about how do we build something that is really going to be the you know the best possible customer
26:55experience and and that that that was a far more effective selling tool than than having a giant sales force or thinking of you know marketing gimmicks or you know 12-step processes or whatever I'm not super familiar with frster certainly I mean the essence of viral marketing is do you have something that One customer is going where One customer is going to sell another customer without you having to do anything and there there are lots of instances of that frienda might be one obviously hot mail was one PayPal was one eBay was one
27:34and and in a situation like that going back to what I said about product product matters incredibly because if you're going to recommend something to somebody you've got to really love the product experience otherwise you're not going to recommend it because you don't want to burn your friend hurdles in the space industry entering the space industry well you know it it is a very complicated regulatory structure as you might imagine when somebody tries to build an orbital launch vehicle which is not really all that distinguishable from an ICBM there's a lot of Regulation and there probably should be because you know you don't want to launch something and end up hitting La
28:15I so probably the regulatory stuff was very difficult the environmental approvals certainly proven very difficult much more so than we respected I mean here in Silicon Valley you look you know what I came to appreciate is in Silicon Valley you live in a Libertarian Paradise you know there's almost no regulation and
28:40what can be very frustrating is that regulation is often irrational it doesn't make any sense but you've got somebody there who's simply executing a set of rules independent of where those rules make sense and you can try to convince them that the rules don't make sense and they won't listen to you so probably regulation is the most annoying thing I would say overall though I'm very pleased because I think we've had a very smooth development process and on the whole I can't complain at
29:09all yeah why is it so expensive to send something into space well let me tell you what makes a rocket hard the the energy and the velocity required to get into orbit is is so substantial that compared to say a car or in a plane the you have almost no margin to play with typically a launch vehicle will get about 2% of its liftoff Mass to orbit so and that's the
29:40case for for Falcon and so if you can only get two% of what your rocket weighs to begin with to to orbit you you can if you're wrong by 2% you're not going to get anything to orbit you know come crashing down on the Pacific somewhere so that means all of your calculations have to be right if if you miss if you miscalculate something you get an answer wrong it blows up and and it's very expensive trying to get all your answers right and then double-checking that they're right and testing them all and doing as much as you can on the ground I think that's a lot of what makes Rockets expensive the low launch rate typically is also what makes rocket Rockets expensive if
30:24you had you know thousands of flights a year then it would be a lot cheaper although it's a bit of a chicken and egg because it needs to be cheaper in order to have thousands of flights a year but at the end of the day when you in the final analysis I would say that Rockets really should be a lot cheaper than they are today and and I think the way they're built the way they're operated is just very inefficiently and I think with with SpaceX and Falcon we're going to we're going to show that that's that's the case our vehicle will sell for about $6 million a flight that our nearest competitor is the Pegasus from orbital Sciences which is about $25 million a flight and that has
31:06less capability than our rocket so Falcon would represent a pretty substantial breakthrough on the cost of access to space can you talk a little bit about difference in the customer base you have to Target experiences ch yeah the customer base with SpaceX is just it's dramatically different obviously from PayPal PayPal is a consumer product whereas SpaceX we're selling Rockets and the number of people who want to buy Rockets is quite small if anyone here has $6 million and wants a rocket I mean I'd be glad to sell it to them but so it's much more of an individual selling process it there's a great deal more thought that goes into any purchase of a launch
31:57much more so than signing up a PayPal account which doesn't really cost you much so yeah and there's not a lot of viral marketing that's going to happen with the rocket I suspect I'm hoping but you know I'm not counting on it so I think a successful entrepreneurs probably come in all sizes shapes and flavors I'm not sure there's any one one particular thing for me you know some of the things I've described already I think are very important I think really U an an obsessive
32:32nature with respect to the quality of the product is very important so you know being obsessive compulsive is is a good thing in this context really really liking what you do what whatever area that you get into given that you know even if you're if you're the best the best there's always a chance of failure so I think it's important that you really like whatever you're doing if if you don't like it life is too short you know I'd say if
33:06and and also if you if you like what you're doing you think about it even when you're not working I mean it'll just it's it's something that your mind is drawn to and and if you don't like it you just really can't make it work I think SpaceX is about 30 people and and what we do internally at SpaceX is we do all of the design analysis integration of Hardware testing and then launch operations but a lot of the heavy Manpower stuff like welding together our primary structure the heavy Machining and so forth that we Outsource so we'd be a much larger company if we did all of that internally and sorry you had another part to your question oh yeah actually we don't have any lawyers
33:58the the the regulatory stuff that we we deal with is very technical it's it's it's really a lot like trying to get an airplane certified with the FAA we're just getting a rocket certified so so our documentation is it's very I I wish we could offload it to to some lawyers they wouldn't know what the heck to do so you know how did the Wharton degree help I think it a business degree teaches you a lot of the terminology it introduces you to Concepts that you would otherwise you know this terminology there's something to be said for that you introduces you to Concepts you you would otherwise have to learn empirically I mean I think I think you can you can learn whatever you need to
34:42do to start a successful business either in school or out of school School in you know in theory should help accelerate that process and I think it you know it often times it does it it's it can be an efficient learning process more efficient than empirically learning lessons but really I think you I there were examples of of successful entrepreneurs who never graduated high school and there are those that have phds so I think the
35:12important principle is to be dedicated to learning what you need to know whether that is in school or empirically well I should point out that Falcon our first vehicle doesn't really have the same capability as as either the Chinese the Russian or or the space shuttle vehicles that I mentioned Falcon would be in the in the light class of launch Vehicles whereas the space shuttle would be a heavy class launch vehicle so it's not it's not quite an Apples to Apples comparison however the right comparison would be Falcon compared to the Pegasus from overall Sciences Falcon is 6 million the Pegasus is 25 million and the way we we've gotten our prices low our cost low is we've
36:04really focused on every element of the launch vehicle there's really no one Silver Bullet that has has been responsible for a substantial portion of of the cost savings it's been really hundreds of small Innovations and improvements and and so we've done improvements in the propulsion system the structure the avionics and the launch operations as well as maintain a very low overhead organization and when you add up all the things we've done in those areas that allows us to produce a launch vehicle at $6 million the as as far as PayPal
36:39there were a lot of back office relationships that we needed to establish and to attach to various heterogeneous data sources we needed to attach to the credit card system for processing credit cards we needed to attach to the Federal Reserve System for doing electronic funds transfers we needed to attach to various fraud databases to run fraud checks there
37:04there was a lot that we had to we had to interface with and and that that that took that took a while but it all came together I think roughly simultaneously I mean developing the software and having it ready for for the general public reasonably coincided with us being able to conclude with those deals and interface with the outside vendors and all that took about a year I think one one thing that's important is to try not to serialize dependencies so if you can put as many elements in parallel as possible a lot of things have a gestation period and there's really nothing you can do to accelerate I mean there's it's very hard to accelerate that just gation period
37:53so if you can sort of have all those things gestating in parallel then then that is one way to substantially accelerate your timeline I think people tend to serialize things too much we did do a few patents on pay on the PayPal system although nothing that ever actually mattered so
38:16patents are mostly useful in in a defensive situation rather than offensive it can be very difficult to actually offensively prosecute a patent so I'm I think I think in certain industries like Pharmaceuticals and so forth patents can be incredibly important in software particularly when you've got a very rapid life cycle where you know you made you sure you got a patent but now it's redundant so who cares it's less relevant in when you've got a rapid life
38:50cycle well I don't you know I've looked at the various there are a couple things that I think are pretty bogus one is space mining another is space solar power I mean if you if you calculate how much it costs to bring either the the photons from space solar power U you know back to Earth or or or the raw material back to Earth it it the economics don't make sense they just can't close the economic case not not even by it's probably off by three orders magnitude so I I think
39:25the probably the biggest thing can happen is if we if we decide to establish a base on the moon or a base on Mars and particularly if we if we attempt to make a self-sustaining self- sustaining base self-sustaining civilization on on the moon or Mars that that that is enormous opportunity on probably the trillion dollar level because then you got basically interplanetary commos going on I think that that's pretty huge but but it's it's it's you know it's not going to be space all power it's not going to be space mining I think so let's see the government I you know who knows maybe we're being spied upon I don't know but but certainly we are there are some
40:15restrictions which are really annoying such as the fact we are only allowed to employ people who have at least a green card or a citizen of of the us we're not allowed to employ anyone who is does not have perent residency in the US us basically if they can't throw you in jail they W they won't let you work work on Rocket stuff and
40:35we have if we talk to any foreign Nationals we we need a a technology transfer agreement or something like that from the state department so our second launch is is actually a non- US governmental launch and it's taken us six months to get the state upon approval just to engage in a contract discussion with them so that that is
41:02problematic we had several offers actually from a number of different entities for for payal and in fact the closer we got to IPO the more more often as we got but we always felt that those undervalued the the company and subsequently when we went public I think the public mon is kind of indicated the value of the company so I think that that's one of the good things about the public markets is that they're an objective value of companies when you're a private company it's very hard to say how much you're worth because you you have to basically think of some Metric you know are you going to go for a multiple of future earnings are you going to go off
41:42some something with Revenue what are your comparables going to be there are all sorts of questions it's it's really up for debate what sort of value your company is when you're public it's it's you know it's whatever the market says you're you're worth that's what you're worth so so so yeah so I think
42:01eBay made a number of office prior to going public that were substantially below our the value once we impose public and that kind of cleared up the disagreement and then we we sold them and what else I was actually you had a second part to the question yeah eBay had a a I it was initially Billo and there was eBay payments and and it was really a pretty tough long running Battle of PayPal versus versus eBay's payment system and it was certainly a very challenging I mean I think it was you know there were times where it felt like we're trying to win a land war in Asia and you know they kind of set the ground rules or trying to beat Microsoft in their own operating system
42:55it's really really pretty hard that took a lot of our effort to to actually beat eBay on their own system and that one of the long-term risks certainly for the company was that eBay would one day Prevail and one way to retire that risk obviously was to sell to eBay writing software during the summer of 95 trying to make useful things happen on the internet and I wrote something that allowed you to get maps and directions on the internet and then something that allowed you to do online manipulation of content kind of a really Advanced blogging system and and then we we we started talking to to small newspapers and media companies and so forth and we
43:50started getting some interest I mean half the time be like what's the internet even in Silicon Valley but then occasionally somebody would and they would would get a little bit of money from them and then U there was basically only about six of us there were myself my brother who I convinced to come down from Canada and a friend of my mom's so and then and then three sales people we hired on contingency by putting an ad in the newspaper but things were pretty tough in the in the early going I didn't have any money in fact I had negative money I had huge student debts so or in fact I couldn't afford a place to stay and an office so I rented an
44:33office instead because that was I was actually I got a cheaper office than I could get a get a place to stay and then we I just sort of slept on the futon and showered at the YMCA on page M and El Camino I was in the best shape I I've ever been you know go to shower work out and you're good to go and and there was an there was an ISP on the floor below us just to like little tiny ISP and we drove the hole through the floor and connected a null modem cable that that gave us our internet contivity for like 100 bucks a month so I mean we had just an absurdly tiny burn rate and and and we also had you know a really tiny Revenue stream but we actually had more Revenue than we
45:18had expenses so so when we went and talked to VCS we could actually well like I said it's there's no Silver Bullet that I can point to as to why the our vehicle is a lot cheaper we've really focused on reducing the cost across the board I mean one thing our overhead in a 30 person company is you know an order of magnitude less than it is in lock or Boeing just just for starters so even if we did everything the same and built the same launch vehicle we'd be considerably cheaper and then every decision we've made has been with consideration to simplicity and the reason for Simplicity is because that both improves the reliability as well as U reduces your cost if
46:03you've got fewer components that's fewer components to go wrong and fewer components to bind I think there's there's I think fairly significant innovation in our airframe which is a a semi- press stabilized monoco with variable skin thickness and a common bulkhead if you know what that means I need a diagram to explain it all but the net result is that it's very cheap cheap and it's very Mass efficient and I think easy to test and quite reliable our avionic system give you another example our avionic system we use an Ethernet on the vehicle to communicate they may that may not sound like a great Innovation but it is in launch vehicles all the other
46:45launch Vehicles communicate in the vehicle by these serial cables that run the entire length of the vehicle and so you've got these these giant copper bundles as thick as your arm running up and down the vehicle makes it heavy it makes it expensive and this so this there's things like that which when you add them all up it it makes a huge difference no that's that's a good question that's a good question no I would not I think SpaceX is not this is Advanced entrepreneuring and it it it may also you know I can't tell you how many people have said that you know the fastest way to turn you the faster way to make a small fortune in the Aerospace industry is to
47:31start with a large one so you know hopefully that doesn't doesn't work out but I I would definit I think space is is a tough one for first-time entrepreneurs you're better off starting with something that that requires low capital and space is a high Capital effort sorry last question
资源与出处
- 视频在 YouTube 收看
- 视频在 B 站收看UP 主「Web3天空之城」搬运,标题注明带中文字幕
- 档案silicon-valley.tech 档案页本页视频与文稿数据来自 Elon Musk Archive
- 精读本篇精读 Markdown 版