大会演讲·中文精读

马斯克在耶鲁大学 2010 完整讲座

Elon Musk at Yale University 2010 (FULL LECTURE)

中文精读由 AI 基于原始视频与文稿改写,非官方翻译,可能存在偏差或错误;表述以原视频为准。说明

2010年,马斯克在耶鲁大学做了一场讲座。他给出的数字是:让人类成为多行星物种,大概要花掉全球GDP的0.25%到0.5%。他顺手给了一个参照系,这笔钱比全世界花在医疗上的少得多,比花在口红上的多一点。

这个比较方式很马斯克。他不谈情怀,先算账。

整场讲座分两块。前半段讲为什么值得去火星,后半段讲SpaceX当时在做什么,以及一个当时还没解决的问题:火箭为什么不能像飞机一样重复使用。问答环节里,他顺带聊了Tesla的续航焦虑、同时运营两家公司怎么睡觉,还有他挑项目的标准。

这场讲座距今十几年,里面有些判断已经被验证,有些还在验证中。但那个最核心的追问没过时:一件事情的优先级,应该由什么来决定?

判断重要性,先看你把时间轴拉多长

马斯克开场没有直接讲火星,他先抛出一个方法论上的问题:你怎么判断一件事重不重要?

他的答案是把镜头拉远。你看的时间跨度越大,当下显得重要的事情就越会褪色,真正重要的东西会留下来。

用这个镜头去看四十亿年的生命史,重要的节点大概有这么几个:单细胞生命出现、多细胞生命出现、植物和动物分化、生命从海洋爬上陆地、哺乳动物出现、意识出现。

他在这张清单上加了一项:生命从一颗行星扩展到多颗行星。

在他看来,这件事的分量和生命从海洋走上陆地是同一个量级。海洋到陆地改变的是生命占据的空间,一颗行星到多颗行星改变的也是这个。区别在难度。后者没法指望几十亿年的自然选择慢慢磨出来,它需要意识参与。

马斯克给了一个很直接的推理:你没法指望生命靠达尔文式的方式,自己演化出跨越几亿英里辐射空间、抵达一个完全不同的环境还活下来的能力。这段旅程必须被设计出来。设计需要意识。

所以在他的逻辑链里,意识的出现和走向多行星是同一件事的两个环节。前者是后者的前提条件。

他并不认为地球快完蛋了

这里有个容易被误读的地方。提出星际移民的人,往往被理解成对地球悲观。马斯克当场把这一点否掉了。

我对地球上的生命其实相当乐观,它大概率会持续很长时间。(马斯克)

他承认存在自然灾难或人为灾难导致生命被抹掉的可能性,但那是小概率。主流情形是地球生命会好好活下去。

既然如此,为什么还要花钱去火星?

他给的理由跟避险关系不大,更多是看到一扇刚打开的窗口。四十亿年来第一次,多行星生存进入了可能性范围。一个从来没被打开过的窗口现在打开了,哪怕你判断地球一切正常,伸手去够一下也是划算的。

资源投入多少,他给了一个可以讨论的区间。不是全部身家,也不是零,大概是GDP的四分之一个百分点到半个百分点,也就是千分之二点五到千分之五。他用口红和医疗做上下界,这个比喻被现场的人笑了,但他显然是认真的:这个数字一出来,事情就从一个空想变成了可以放进预算表讨论的量级。

到太空和到轨道,是两回事

讲SpaceX之前,马斯克先做了一次科普,因为这两个概念经常被混在一起。

很多人以为SpaceX和理查德·布兰森的维珍银河在竞争。他说不是。维珍银河只是到太空边缘,需要的终端速度大概3马赫。而进入地球轨道需要的终端速度是25马赫。

差别的大头在能量。动能和速度的平方成正比。25马赫和3马赫之间差的不止七倍速度,能量差了大约五十倍。算下来,仅仅“到太空”只需要“到轨道”所需能量的2%。

这个数量级的差别解释了为什么维珍银河那种方案在工程上可行,而进入轨道难得多。马斯克还顺手纠正了一个常见误解:两三百英里高度上的重力,和地面上的重力几乎一样强。飞行器之所以不掉下来,是因为它横向速度足够快,向外的加速度正好抵消了重力的向内加速度。失重感来自这两个加速度相消,跟重力消失无关。

前三次都炸了,第四次才活下来

Falcon 1是SpaceX的第一枚火箭。马斯克给了一个直观的类比:如果它是飞机,大概相当于湾流5那种公务机的尺寸。

然后他讲了四次发射的故事,语气平静,内容不太平静。

第一次发射,升空60秒后发动机起火,发动机停机,任务失败。

第二次,飞得相当接近轨道了,但二级出现控制问题。他在这里解释了一个结构性难题:二级发动机是在真空中、高速、高过载条件下点火的,这个状态在地面上没法真实测试,只能仿真。而他们的仿真“恰好过于乐观”。

第三次,他们改主发动机设计,从烧蚀冷却换成再生冷却。烧蚀冷却是让燃烧室壁慢慢烧掉,再生冷却是把燃料通进燃烧室壁里带走热量,壁不烧掉。改完之后出现了一个推力瞬变,时间比预期长,导致级间分离失败。

第四次,成功入轨。

马斯克没有掩饰那一次的压力:

如果第四次还没成功,SpaceX今天是否还存在,就不好说了。(马斯克)

这句话的分量在于,一家公司把生死压在一次发射上的时候,前面那三次失败就变成了必需品。他后来在回答提问时说,如果自己当年对火箭懂得更多,也许不用等到第四次就能入轨。这句自嘲里有一层意思:外行进技术项目,先交的是学费。

第五次发射也成功了。之后NASA给了合同,让SpaceX成为航天飞机货运的主要替代方案。他提到,按当年的进展看,载人运输的主合同大概率也会拿到。

Falcon 9首飞成功,因为学费已经交过了

Falcon 9比Falcon 1大得多,名字来自底部那九台发动机。马斯克提到,每一台发动机的推力都比Falcon 1的发动机高,整枚火箭的推力超过一百万磅。他先说这是波音747最大推力的四倍,随即自己纠正,说大致相当。

2010年6月,Falcon 9首飞成功。火箭顶部装的是Dragon飞船的结构测试件。马斯克说了一句挺得意的话:看起来我们该学的都学到了,所以第一次就成功了。

Falcon 1从马绍尔群岛的夸贾林环礁发射,Falcon 9从卡纳维拉尔角发射。他放了几段视频,其中一段是配上电子音乐《High Roller》的发射剪辑。

关于Dragon,他讲了一个当时还没人做过的设计。传统的发射逃逸系统是在飞船鼻锥上装一根牵引塔,出事时把飞船拽走。SpaceX的方案是把逃逸系统直接做进飞船本身,鼻锥上没有那根塔。他的判断是,这样从根本上更安全,也更高效。

还有着陆方式。短期内Dragon溅落在海上,长期会落在陆地上,把着陆腿弹出来,落在加州莫哈韦沙漠一带。他顺带讲了再入的精度问题:飞船带着升力再入,靠船上发动机转动飞船就能改变升力方向,从而控制落点。阿波罗时代没有GPS,落点精度已经能到一英里以内。

真正缺的那个发明:完全可复用的火箭

讲座后半段进入核心。马斯克说,让生命成为多行星物种,需要的根本性突破只有一个:完全可复用的轨道级火箭系统。

为什么这件事这么重要,他给了一道算术题。一枚火箭造价一亿美元。用一次,每次飞行成本就是一亿美元。用一千次,每次十万美元。差距是三个数量级。

然后他扩展到所有交通工具。任何交通工具如果不可复用,几乎就没有使用价值。自行车、马、飞机都一样。你骑一趟马,回程还得再养一匹马。

飞机是最好的例子。一架波音747造价两亿五到三亿美元,往返一次需要两架。但你提前订票,从纽约飞伦敦可以低于一千美元。原因就是这笔资本成本被分摊在一两万次飞行上。

航天飞机是唯一的反例参照。它部分可复用,主燃料箱每次扔掉,剩下那些号称可复用的部件,翻新起来极其麻烦。结果航天飞机每次飞行的成本,是同等级一次性运载器的四倍。马斯克的原话是,它甚至是唯一一个还在运行的、勉强算部分可复用的例子。

他还讲了一句挺刻薄的话:历史上有很多次造可复用火箭的尝试,最后都被取消了,因为立项之后大家发现,成功不在可能的结果列表里。

SpaceX当时在做什么?马斯克很诚实:除了捡回一些碎片,什么都没回收过。他希望下一次Falcon 9加Dragon的飞行,如果一切顺利,能回收一级火箭和Dragon飞船。再往后是回收二级,再往后是把整个回收流程做到成本上真的划算。

他说,不知道能不能成,但必须试。

火箭贵不贵,在图纸阶段就定了

问答环节第一个问题问的是竞争格局。马斯克给了几组数据。

Falcon 9的运力大约是12吨到轨道,和洛克希德、波音的同类火箭相当,或者说相当于RN5两个助推中的一个。价格上,SpaceX大约是波音或洛克希德的三分之一到四分之一,是俄罗斯或中国报价的一半左右。他说这是当时全球成本最低的。

为什么能便宜这么多?他把功劳归到设计上,跟劳动力便宜没多大关系。

系统的设计决定了潜在成本的下限。如果你设计得足够聪明,就能抵消其他国家的劳动力成本优势。(马斯克)

他还强调了一个常被忽略的做法:把工程和生产放在同一个地方。这样设计和制造之间的迭代、优化可以很快发生。在他看来,这件事的价值被严重低估了。

这个观点值得单独拎出来。很多人算制造业成本,算的是人工、土地、税收。马斯克算的是“改一次设计要多久”。一个改动如果需要跨时区沟通两周才能进产线,那这个组织的迭代速度就被锁死了。省下来的人工费,抵不过慢下来的损失。

续航焦虑,其实是个平方问题

第二个问题问电动车的续航。马斯克当时在准备Model S,给的方案是三条腿走路:续航做到300英里,快充45分钟,以及快换电池包,换包速度比加满一箱油还快。

然后他讲了一个很漂亮的数学关系。

对充电站和换电站的需求,随续航的平方下降。(马斯克)

为什么是平方?因为续航决定的是覆盖面积,而面积和半径是平方关系。续航翻倍,你能覆盖的面积是四倍,需要补能的频次相应下降。

按这个逻辑推下去,300英里续航已经不需要在城市里布充电站了,只需要在州际公路沿线每150到200英里放一个。如果续航做到500英里,几乎就不需要公共充电站了,除了你住的那家酒店。因为一天开500英里以上的人,少到可以忽略。

这个推理方式很典型。他没有去争论“充电桩够不够”,他换了个角度:把续航这个变量调大,需求本身会以更快的速度消失。解决一个问题的最好方式,有时候是让这个问题不存在。

100小时工作周是怎么来的

有人问,同时管SpaceX和Tesla,时间怎么分配,睡觉怎么办。

马斯克的回答挺坦白。他本来没打算管两家公司。他的计划是自己管SpaceX,Tesla交给别人。结果金融危机来了,汽车行业受冲击最重,通用和克莱斯勒都破产了。

摆在面前的选择有两个:要么少睡觉、接手Tesla、把剩下的资本全投进去;要么Tesla死掉。

他选了前者,代价是那段时间连续每周工作100小时。他自己评价那段日子:

那段时间真的不好玩。(马斯克)

到2010年讲这话的时候,他说情况好多了。两家公司的管理团队都更有厚度,他能睡个合理的觉了。

这段回答里有个细节值得留意。他没有把“每周100小时”包装成一种值得效仿的工作伦理。他把它描述成一个被迫的选择,一个特定情境下的应急状态,而且明确说不好玩、不享受。这和后来很多把加班浪漫化的叙事是两回事。

从PayPal到造火箭,中间隔着一次彻底的外行

有人问他,从互联网创业者转型做火箭,这个跨越是怎么完成的。

马斯克先回答了“是什么把这些公司串起来”。他说大学时他就在想,哪些东西最影响人类的未来。他想出来三个:互联网、可持续能源的生产和消费、太空旅行。SpaceX、Tesla、SolarCity、PayPal,只是这三个方向的具体形式。

然后是转型的具体过程。他说得挺直接:在PayPal之前他只做过互联网软件,从来没造过任何物理的东西。那几年基本是从零学起,跨度极大。

最初他打算雇一个人来做火箭的总设计师。结果找不到。

能胜任的人不愿意来,愿意来的人干不了。所以我最后成了火箭的总设计师。(马斯克)

他还补了一句:如果当年他对火箭懂得更多,也许不用等到第四次发射才入轨。

这句话有两层。一层是自嘲,说明外行领导技术项目要交学费。另一层是,他承认自己当时不懂,但项目还是往前走了。这两件事同时成立,才是这段经历的真实质感。

NASA该做什么,私营公司该做什么

有人问NASA和私营部门的边界。马斯克的回答分两半。

前半段是感谢。NASA是SpaceX最大的客户,当时卖出去的发射任务里,一半来自NASA,一半来自其他机构。NASA还在一些关键技术问题上充当顾问。SpaceX本身也是站在早期航天计划积累的知识上往前走的。

后半段是批评。他的判断很直接:政府总体上不擅长创新和成本优化。

所以他给的分工是:NASA定目标,说清楚需要达成什么,然后让私营部门去找解法。航天飞机退役后的货运补给就是这么做的,12次任务给SpaceX,8次给Orbital Sciences。他认为这比之前的设计好得多。

但NASA的角色不能取消。马斯克把它比作一个太空领域的风险投资人:出钱支持那些对整个国家、甚至对全人类有好处,但没有直接商业价值的项目。送火星车上去、造哈勃望远镜,这些东西的价值很难折算成某家公司的收入,所以必须由公共资金来承担。

他还提到另一类该由NASA做的事:基础研究,尤其是那些回报周期特别长的方向。

他挑项目的标准,不是投资回报率

有人问,你选项目的顺序是先看有没有意思,还是先看有没有商业潜力。

马斯克的回答是,两个都不是。他用了一个词:相关性优化。先问这件事对人类未来的影响是不是最大,然后再去看产出价值能不能大于投入成本。后者是必要条件,任何持续经营的企业都得满足。

他明确说,如果按投资回报率排序,太空和汽车会排在很后面。这两个领域的回报极难做出来,他自己做成了,但过程极其艰难。相比之下,再做一家互联网公司,回报会来得容易得多。

那为什么还做?因为影响力。而把经济账做平,是获得影响力的入场券。

如果你不把经济账做平,你就没法对未来产生任何影响。(马斯克)

这个逻辑闭环挺重要。很多人把“改变世界”和“赚钱”对立起来,马斯克的处理方式是把两者排成先后:影响力是想要的东西,经济账是能继续做下去的前提。前提不成立,想要的东西就无从谈起。

他为什么不肯卖公司

有人问起Zip2和PayPal的退出。马斯克说他其实一直不太想卖。Zip2是他第一家公司,做的是帮媒体公司上网的软件。PayPal那次,他是最不想卖的那个人,但董事会其他人想卖,他就跟着走了。

然后他把话题转到现在:Tesla、SolarCity、SpaceX,都没有卖掉的打算。

SpaceX尤其不能卖。他的理由是,一旦卖掉,公司就会失去他设定的那套哲学目标。

Tesla是上市公司,理论上存在被恶意收购的可能。但他持有大约三分之一的股份,所以对方要成功,前提是他自己做了非常愚蠢的事。他的原话是,只要他做的事只是“中等程度的愚蠢”,收购就不会成。

SpaceX未来可能会上市,但他会保持多数股权。上市图的是拿到低成本资本,控制权他不打算放手。

Tesla呢?他说短期到中期不会卖,但未来某个时点,如果股东想卖,他可能会同意,前提是Tesla能实现的根本价值已经基本实现。

那个根本价值是什么,他讲得很清楚:

我对Tesla最好的期望,是它的存在把电动车的转型加速了5到10年。(马斯克)

为什么加速这几年很重要?他给的理由是公地悲剧。生产二氧化碳的过程在消耗海洋和大气吸收二氧化碳的能力,但没人在加油站为这部分买单。这个缺口没法靠市场自己补上,只能靠创新去填。Tesla和SolarCity的目标就是填这个缺口。

给年轻人的建议:别从火箭和汽车起步

有个学生问,作为学生,做能源和太空比做互联网难太多,有没有什么建议。

马斯克的回答很务实。如果你刚起步,不管是做工程师还是创业,应该选资本需求低的方向。互联网之所以是好选择,就是因为这个。

另一个方向是,如果你能在风能或太阳能的效率上做出真正的发明,那有可能吸引到风险投资。

但太空和汽车,他的说法是:

太空和汽车大概相当于研究生级别的创业,从这两个领域起步并不明智,资本门槛太高了。(马斯克)

这个建议和他自己的路径是矛盾的,他自己就是从这两个领域起步的。但他没有因此修改建议。他知道自己是个例外,而例外不该被当成模板。

可复用火箭到底难在哪:2%这条线

最后一个技术问题问得最实在:造可复用航天器的主要障碍是什么,SpaceX打算怎么跨过去。

马斯克的回答揭开了整个问题的底层约束。

一枚一次性的火箭,用最好的材料、最好的制造工艺,最后能把起飞质量的2%到3%送进轨道。这个数字就是天花板。

一旦加上回收系统、热盾、各种为了回来准备的东西,重量都要从那2%到3%里扣。

历史上那些可复用火箭项目,扣完之后发现,入轨载荷变成了零,或者接近零。这就是它们被取消的真实原因。技术本身没到瓶颈,算术先撑不住了。

所以真正的挑战是三件事同时做到:发动机效率更高、结构质量效率更高、回收系统足够轻。目标是扣掉所有回收系统之后,还能剩下至少2%的起飞质量作为载荷送进轨道。

SpaceX给Falcon 9第二版设计的方案是:一级火箭反推返回,在发射场动力着陆;二级用翼伞,最后靠气囊或者某种软着陆方式落地。按设计,这个方案理论上能保住2%的载荷,代价是入轨能力降到原来的十分之一左右。

马斯克强调,这里没有银弹。

这里没有某一个点的突破,全靠整个系统的优化。发动机、主结构、航电、回收系统,全都要一起算。(马斯克)

SolarCity卖的不是太阳能板

最后一个问题关于SolarCity。马斯克先纠正了一个常见误解:SolarCity是当时美国最大的光伏太阳能供应商,但它不生产太阳能板。

他给的类比是戴尔。戴尔不造硬盘,不造CPU,它做的是把部件组装成一台适合你的机器。SolarCity做的是为某个具体的屋顶设计整套系统,负责安装、布线、逆变器、审批许可这些事。

为什么这件事值得单独做一家公司?因为成本结构。在屋顶装太阳能,面板只占40%的成本,剩下60%是别的。那60%就是所谓的“系统平衡”,是最琐碎、最难标准化、也最容易被浪费的部分。SolarCity的价值就在这60%里。

面板本身,他说美国公司其实很强。First Solar在成本上做得非常出色,SunPower在转换效率上最高。大量的创新发生在美国。

把这场讲座拆成可以用的东西

十几年过去,这场讲座里有些东西已经被验证,有些还在路上。Falcon 9后来真的实现了一级回收和复用,猎鹰重型飞了,Dragon开始载人,星舰还在试。Tesla把电动车从边缘产品变成了主流选项,加速的年份大概对得上他说的5到10年。

但真正值得带走的,是他处理问题的那套方法,结果反倒次要。下面这几条,普通人也能直接用。

  • 判断一件事值不值得做,先把时间轴拉长。问自己:十年后回看,这件事还会显得重要吗?如果答案是会,那当下那些让你犹豫的噪音,多半可以忽略。
  • 算清楚分母再谈投入。马斯克把星际移民的成本定在GDP的千分之几,这个数字一出来,讨论就从“值不值得”变成了“划不划算”。任何宏大的想法,先换算成一个可以放进预算表的数字。
  • 遇到瓶颈,先检查是不是算术不成立。可复用火箭卡了几十年,卡住它的东西是那2%的质量预算。如果你推进了很久的事情一直推不动,试着把约束条件写成一道算式,看看是不是某一项根本对不上。
  • 把工程和生产放在同一个房间。如果你在带团队,缩短“改一版设计”到“看到结果”的时间,比省下多少人力成本都重要。
  • 面对一个需求,先想能不能让它消失。续航焦虑不靠多建充电桩解决,靠把续航做长,让需求按平方衰减。做产品时,先问这个需求是不是可以用另一种方式绕过去。
  • 分清目的和条件。影响力是想要的东西,经济账是能继续做下去的前提。前提不满足,目的就是空话。别把两者混为一谈,也别因为看不起条件而放弃目的。
  • 起步阶段选资本门槛低的方向。这是马斯克给学生的原话,虽然他本人是反例。反例的价值在于说明可能性,不构成建议。

最后一条可能是这场讲座里最不像马斯克、但最实在的话。一个从火箭和汽车起步的人,回头告诉年轻人别从这两个领域起步。他知道自己走了一条不该被复制的路。

YouTube 平台字幕(自动或人工),可能有识别错误,人名与数字请以视频为准。已按语句合并成段落,并去掉 um、uh 等口头填充词。 文稿来源:YouTube(自动/人工字幕)。

0:00Great pleasure to be here. Thanks for Thanks for having me. the the what I was going to talk about today was basically the the importance of making life multi-planetary and kind of how I think we could make get there and it's I think it's it's one of the most significant things that that will ever occur in in the history of life itself and and this is really the first time in the 4-billion-year history of life on Earth that it's been with with in the realm of possibility.

0:33You know, first of all, I guess one should sort of look at at the the nature of importance itself. How do you decide that that anything is important? and I think that the lens of history is is a helpful guide in that the further you zoom out and the greater the span of time that you look at, the things that seem important in the moment fade away and the the things that are truly significant are are still there. And if if you look at the the history of the evolution of life, you could say, "Well, what are the major elements in that in that history?" And then obviously the advent of single-celled life, multicellular life, differentiation of plants and

1:13animals, the transition from of from the oceans to land, mammals, and consciousness. I think those are probably the the big ones, but I think also on that list would would fit life going from one planet to multiple planets. you know, it'd be sort of at least comparable to life going from the the oceans to land.

1:34And and I think given the the immense difficulty of of of of that, I think it we're you you actually need consciousness to develop in order to to achieve that goal because this I can't sort of see any way that life could have could actually just evolve in Darwinian fashion to go across hundreds of millions of miles of irradiated space to an environment that's completely different to it, and still still live.

2:06So, you really need consciousness in order to design a mechanism of of making that journey, and then I don't know if I should Actually, I'll just stand. Talking to a mic and I've got a thing on. so, anyway, so it's I think an extremely important thing that we we we try to accomplish this, and it's not really from the perspective of of thinking that the life on Earth is in in imminent danger of of of ending.

2:40I'm actually quite optimistic about life on Earth. I think it's likely to to last for a long time. There's this this there's some chance of of life you know, being being eradicated due to some natural calamity or potentially a man-made calamity. But, I I'm I think most likely it will will do well for a long time.

3:02nonetheless, this is the the first time that it's been possible to consider making life multi-planetary, and so I think it's it's wise to take advantage of that opportunity, and and dedicate some portion of our resources of our collective resources to that goal. and we can debate as to what is the magnitude, how much you know, resources should be dedicated to it.

3:31but, it's you know, I think not nearly as much as we dedicate to to health care, but maybe more than we spend on lipstick. You know, it's true. That's that kind of a brackets thing. I don't have anything against cosmetics. I think they're great. but, you know, it just that's it it you know, gives you an approximate sense of things. Maybe it's a a quarter of a percentage point of the GDP or half a percentage point of GDP or something like that.

4:04So, in in in doing SpaceX, what what what we're trying to do at at SpaceX, we're a rocket company, is to to make some progress in that direction. And it's it's a very difficult goal, so I think it's going to take a lot of companies and a lot of people to to actually make this happen, but at SpaceX, we're going to do our best to move the world forward as much as we can.

4:32So, the the way we started out was by designing a small rocket called Falcon 1. and if that was an aircraft, it would be comparable to like a a Gulfstream 5 or something like that. Basically, comparable in size to a a corporate jet. And it's It it took us a couple tries to get to orbit. It's quite difficult.

5:05Getting to space is actually a lot easier than getting to orbit. So, sometimes people think that SpaceX is competing with Virgin Galactic, which is being done by Richard Branson, but but this is not the case. Virgin Galactic is just going to the edge of space, which requires a terminal velocity of about Mach 3.

5:24but getting to orbit requires terminal velocity of Mach 25. and the energy scales with the square of the velocity. So, getting to space is only about 2% of the energy required to get to to Earth orbit. So, the the the first launch that we did with Falcon Falcon 1 didn't didn't even get to space. Got 60 60 seconds off and then an engine fire terminated the well, it's it's a complicated story, but the suffice it to say an engine fire caused the engine to stop. And and then on flight two, we actually got quite close to orbit, but had a a controls issue on our second stage.

6:12it's very difficult to test the second stage in particular because the second stage starts in vacuum at very high speed and and high G loading. So, it's really difficult to test that exactly on Earth. You can simulate it, but you can't really test it effectively. and our simulations happened to be optimistic.

6:35so we we corrected that error and then on flight three, we we changed the the main engine design from an ablatively cooled main engine, which means that the the chamber wall burns away slowly to a regeneratively cooled engine, which means that's that you use the the fuel to cool the chamber. and essentially you run the fuel the fuel through the the chamber wall to cool the chamber and and and so nothing burns away, but the net result was a thrust transient, an extended thrust transient that caused failure on stage separation.

7:15so But fortunately, fourth time's a charm, as it turns out. I think that's how the saying goes. and and on flight four, we we we made it to orbit with Falcon 1. and that was a very nervous time because we were If we hadn't made it on flight four, it's not clear whether we would have whether SpaceX would would still exist today.

7:47and but fortunately we did make it on flight four and and then flight five was also successful. and we were awarded the contract in NASA to be the primary replacement for the space shuttle as far as cargo transport is concerned and with recent developments this year, looks like we will win the primary contract for astronaut transport as well.

8:12and and then we launched the Falcon 9 which is a much larger version a much larger rocket. And and we launched that in June and it it looks like we learned our lessons because it actually was successful on the first flight. so I'll show you some of the videos. at some point we'll maybe release videos of the first three flights.

9:23This is the the music video version of the launch. It's the music we're sampling is Highroller from The Crystal Method.

10:12Okay, in the background there you can see a white rock. That's the little island that we launched from. that's part of the Kwajalein Atoll in the Marshall Islands.

10:315 4 2 1 5 4 2 1 5 4 2 1 5 4 2 1 That was the talk on And if it's really difficult the real difficulty with getting to to orbit is really that that velocity element. Some of these people sort of assume that if you go up to a certain altitude I mean obviously you're engineers so that you you know this isn't true but some of them will think that gravity kind of stops if you if you go up a couple hundred miles. But actually the gravity the strength of gravity a couple hundred miles above the surface of the Earth is almost the same strength as at the surface of the Earth. It's just that in order to stay up there, you have to zip around the Earth really fast

13:42such that the outward acceleration the outward sort of radial acceleration is equal to the inward acceleration of gravity. And so these two null out and that's what gives you the the feeling of zero-g. So here I'll show the Falcon 9.

14:28So the Falcon 9 is a much larger rocket than the Falcon 1. In fact, it's it's named after 9 because it has the nine engines on the base. And each of those engines is actually operating at a higher thrust than the Falcon 1. So you really have something here which is over a million pounds of thrust and that's about four times the maximum thrust of a 747.

14:52let's go let's go wait a minute. It's roughly equal to the 747. This is the launch that took place in June. and it's got essentially a it's just a structural test article of our Dragon spacecraft on the top there. the launch that's coming up hopefully next month will carry an operational version of our Dragon spacecraft.

15:18And this is Cape Canaveral. So Falcon 9 launches from our launch pad we have at Cape Canaveral, whereas Falcon 1 launches from our launch apple and watching apple?

16:41Looks like it's progressing. Don't go trans sonic there. Repeat 5.

17:38Towards the end of the first stage, burning will actually start shutting down engines to be able to do that.

18:19This is the phase of flight that we were most concerned about is the stage separation and then ignition and control of the second stage. In this case, the second stage engine was a turbo pump fed engine. Which is more complex than a pressure fed engine.

19:04It was also rolling a little more than it should. The second stage roll control system was is not functioning quite as well as it could have.

19:25So, that was Falcon 9 Future Lift to see that one get away. then I got one final video. Just watching. Now, this is obviously a simulation. And this this is a simulation of the the Falcon 9 Dragon spacecraft it going to this the space station. And basically the Falcon 9 booster delivers Dragon to a parking orbit near the space station. Then our Dragon spacecraft maneuvers over the space station, docks docks, delivers cargo or astronauts, and then undocks and returns to Earth.

20:07in this simulation, the launch escape system is shown as a kind of a tractor, like that needle needle thing on the nose. but the we're actually going to do something that's and I've never been done before, which is to build the launch escape system into the spacecraft itself. So, you don't have a a tractor on the nose.

20:26we think that's that's fundamentally safer and and more efficient. also Dragon once we've we're confident of our landing accuracy, Dragon will land on land as opposed to the water, which is shown in the simulation.

21:08Flying is on final approach. So, as I mentioned, next month we'll be launching the first operational version of our Dragon spacecraft. It won't dock to the space station on the first flight. It's going to just maneuver around the orbit and then reenter and test the primary heat shield, the guidance, navigation, and control system, all the telemetry, and then the obviously the parachute recovery from from the water.

21:38and then flight the the flight after that, which is due to take place next year, will be the first flight that actually goes to the space station. So, if all goes well, we'll be going to the space station and doing pretty much what you see here, except with cargo instead of people, around the middle of next year.

21:53Which is mass in the capsule. So, just by rotating the capsule with the onboard engines, you can actually change that lift vector and achieve a very accurate landing location. in fact, even during the Apollo era, they were able to land to within an accuracy of about 1 mile. and that was without GPS or anything. So, it it sort of seems like it was coming in ballistically, but actually it's a very controlled reentry.

22:58Hmm. And then the And then as I said that this is kind of a short-term thing long-term will land on land. So we'll just basically pop out landing legs and land somewhere probably in the Mojave Desert in California.

23:45And so this is sort of where SpaceX is at this point that Now the really big breakthrough that's needed the fundamental invention that is needed to make life multi-planetary is a fully reusable orbital orbital class rocket system. This is an extremely difficult thing to achieve. The the space shuttle which is partly reusable I say partly because the main tank is thrown away every time and even the parts that are supposed to be reusable are so difficult to reuse or refurbish that the space shuttle costs about four times as much per flight as an equivalent expendable vehicle.

24:29So and and space shuttle is is the only operating example of something that even has partial reusability. there have been many attempts to achieve to build a reusable rocket system but they have always been cancelled once they discovered that success was not one of the possible outcomes.

24:52So But but this is fundamentally something that has to be solved if if life is to become multi-planetary. and it it it is just incredibly fundamental. if if I mean, if a rocket costs, say, a hundred million dollars to produce, then if you can use it once, well, then it costs a hundred million dollars per flight. But, if you can use it a thousand times, it costs a hundred thousand dollars per flight. It's a pretty huge difference. and that that's the reason why, for example, air flight, okay, you think of any mode of transport, and imagine if it were not reusable, it would see almost no use.

25:29whether it's a bicycle, a horse, plane, I mean, just think, you'd have to have two horses just for a round trip. and you know, or air flight's a good example where, you know, a 747 costs on the order of a two hundred fifty and three hundred million dollars, and you need two of them for a round trip, and yet you can fly from New York to London for less than a thousand dollars if you book your ticket in advance. So, the reason is because the capital cost of that 747 is allocated over Well, in the case of 747, probably ten or twenty thousand trips.

26:07so, that that that's really the the hugely fundamental thing that's needed. and and SpaceX we're we're hoping to make some progress in that direction. thus far, we've not recovered anything except a few little bit bits and pieces. and but hopefully on the upcoming flight of Falcon 9 with Dragon, if things go really well, we'll recover the first stage, and we'll also recover our Dragon spacecraft. And then long term, we'll have to work on recovering the second stage, and then making that whole recovery process really cost effective.

26:43And yeah, so that's that's kind of like the the long term goal. and I don't know if we'll be successful on that, but I think we we have to try. Yeah, so that that's that about SpaceX. And this is the main topic of my talk, but I'm happy to take any questions from people about SpaceX or Tesla or or anything else.

27:24We have plenty of time for questions, but we don't have any mics in the audience. So, if you'd like to ask a question, just make sure you boom it out. We'll start in the back, Sandra. two questions. One on the space part and one on the automotive part on Tesla. On the space part, you just mentioned that the long-term goal is to recoverability.

27:47Currently, how many launches do you have in the pipeline? How many rockets do you have there? You see those of friends and alliances and in terms of performance, cost, and money. Sure. And the second question is on the Tesla part. As you know, people talk all the time about range and the side.

28:06How do you see the future of that when when it comes to electric cars as far as range is concerned? What solution do you see for following that type of thing? Right. Well, there's always Zeniths and Red Jackets. So, well, in terms of the the cost of Falcon 9 capability, Falcon 9 Falcon 9 is capable of about roughly 12 tons to orbit.

28:34and that's comparable to the rockets made by Lockheed and Boeing domestically or comparable to one of the two bays in the RN5. RN5 is a bigger rocket that launches two satellites at a time. and we we actually have the most cost effective in the world. we're about a third to a quarter of the cost of Boeing or Lockheed. And, and maybe half the cost of the Russians or the Chinese.

29:07and I you know, it's I think it's just really a function of of the design of the system that it's enabled to to be a lot more cost effective than than other other designs. and I'm a big believer in having engineering and production co-located so that the the iteration we the optimization between engineering and production can happen quickly.

29:35so I think that's often overlooked as as a way of as a way of achieving cost effective product is is you know, what whatever the design of the system sets the sets the the the the lowest potential cost. and if you so if you're smart about the design you can overcome the the labor cost advantage of other countries.

29:59and then with respect to the range anxiety thing, the way that we're addressing that with the Model S, the sedan that's coming out in a few years, is we're offering a range of up to 300 miles, an ability to charge the battery very quickly, and and recharge it in about 45 minutes, and then also fast pack swap. So, the battery pack can be swapped out faster than you can fill a gas tank.

30:28and if you if you go to a range of 300 miles, you really don't need a a lot of recharge stations or battery pack swap stations. In fact, you don't need any within your hometown. You maybe need them along the interstates and that kind of thing, every 150 to 200 miles. I mean, it's interesting to note that the the need for recharge stations and battery quick pack swap stations drops with the the square of the of the range.

30:54You know, that's cuz that defines the the the area you're able to cover. and so, as the range gets bigger and bigger, the need for recharge stations drops quite a bit. And arguably, if you get up to around 500 miles of range, you don't really need any charge stations because yeah, apart from the one at the hotel that you're staying at. because almost nobody travels more than 500 miles in a day.

31:19Right in the middle. So, as you mentioned earlier, you're working on these two like parallel side projects at the same time, right? So, how do you decide where you allocate your time and your life? Where where does sleep come into it? Great question. well, yeah, it's definitely been a been a difficult thing to to run SpaceX and Tesla simultaneously. And it was not it it was not really my intent to run both companies.

31:47I I had a I hoped to just run SpaceX and then Tesla would be run by by others, but when the economy you know, had issues, that and and those issues particularly affected the car business. I mean, you saw Motors and Chrysler went bankrupt. it was a choice of either get a lot less sleep and and run Tesla as well, plus apply all my remaining capital to Tesla, or Tesla would die.

32:17and so, that's you know, I I obviously did the former. And that that was definitely it was actually definitely not fun during during that period. but at at this point, I feel a lot better about things. and the management teams of both companies are a there's there's a lot more bench strength there, so I can get

32:44a reasonable amount of sleep. But for a while there it was sort of continuous 100-hour weeks, which is I think that's not fun. Go ahead. Hi thanks for your talk. I personally see you as like an amazing inspiration as an entrepreneur. And I realize there's a significant difference between being a successful entrepreneur, which you were with X.com and PayPal, and then tackling an industry as innovative as space travel, and actually rocket science.

33:17so I was wondering if you could speak a little bit about making the transition to space after initially doing that to get there. sure. well, the the the the I should perhaps mention that the thing that kind of ties together SpaceX, Tesla, SolarCity, which we haven't talked much about, and and and then PayPal, and my first internet company was that in in college I thought about what are the things that would most affect the future of humanity, and those things seemed to be the internet, sustainable energy, both production and consumption, and then and then space travel, particularly making life multi-planetary. So that's that's the only thing that really ties

33:59them together. But going from from PayPal to SpaceX was definitely a a huge learning process, because I'd never made anything physical before. I just basically done internet software. so there were it was just a huge huge learning learning process, and the initially I had thought that I would hire someone to be the chief designer of the rocket, but I couldn't actually find anyone who was willing to join and who I who I who was good. So, people who weren't willing to join who I thought weren't couldn't do the job and those who could do the job weren't willing to join. So, I ended up being chief designer of the rocket. and I think perhaps if if

34:44you know, if I'd been more knowledgeable, we would have gotten to orbit sooner than flight four. but it was Yeah, that's kind of how it how how it went. yeah. So, It seems like SpaceX makes like a compelling story for the privatization of space exploration. I'm curious what you think SpaceX and the private sector should be doing versus what NASA and the public sector should be doing. Is there anything NASA should be doing that you shouldn't be and so on and so forth? What are what are the goals?

35:21well, I think NASA certainly deserves a lot of credit for the progress that SpaceX has made in that NASA's our biggest customer. in all of the launches that we've sold, about half are to NASA, about half are to other organizations. NASA has also served as as a an advisor on some important technical issues.

35:42and so, they've been very helpful. And of course, SpaceX is building upon knowledge that was created during the early days of the space program. now that that being said, that I mean, NASA is still at the end of the day, the government. So, the and the government is generally not good at generally not that good at at innovation and cost optimization.

36:07so, it is it's better if if NASA is kind of sets the the goals and says, "Okay, this is this is what we need to achieve." and then allows the private sector to figure out the right solution. and I think that's been very effective with the transition away from the shuttle to SpaceX and then there's another company Orbital Sciences.

36:32NASA awarded 12 12 of the resupply missions to SpaceX and eight to Orbital Sciences. So, I think I guess a much better solution than what is contemplated before. But but I still we still need that that that that NASA there to serve as you know, a funding source essentially kind of like a space venture capitalist if you will.

36:57That's spending money on things that are of benefit to the whole country and arguably to humanity as a whole but aren't necessarily of commercial value need to that that plugs directly to the economy. You know, it's it's hard to say exactly like what is the direct commercial value of sending the rovers to Mars. It's I mean it's or the Hubble telescope. You know, these things are of great value to to humanity but they aren't of specific commercial value to other companies.

37:29And so that there is that important role for for NASA to play in you know, making and in funding those important objectives and perhaps in working on you know, doing basic research on on sort of fundamental new things that that have a very long payoff period. Either. You started a number of really interesting things which are fun and I wanted to ask you what inspiration

38:03do you first look to whether there's a need and an economic potential or if you look at something that you think would be interesting and fun for you today and then see whether there is a possible economic potential and utility and follow but just in terms you to think what comes first with you.

38:30Sure. Well, I'd say it's actually it's it's neither of the two. It's it's really kind of like a relevance optimization. It's What do I think is going to make the biggest difference to the future of humanity and and then and then I see try to see if I can have the the value of the output be greater than the cost of the input.

38:55Which is necessary for any ongoing enterprise. So, but but it's definitely not from perspective of this is the best return on investment or anything like that. If one were to rank order return on investment and the amount of effort required, I think space and and cars would be really low on the list.

39:17It is extraordinarily difficult to make an economic return on you know, on a space endeavor or like I did on a car company. It's It's certainly possible and that's where actually it has turned out to be a good economic return. But but at at great difficulty. And if I'd done say another internet company, the the return would have been it would have been easy relative to the you know, the return that would have it would have yielded.

39:47So, yeah, so it's just from the standpoint of I'm working on things that I think will have the biggest impact on on the future. And then economics the making the economics work is necessary just because if you don't make the economics work, then you don't have any impact on the future.

40:05Wait, is this the second row? Follow up on a question. Just thinking back on your other ventures, your thoughts on things like Zip2 and PayPal. What do you think of those ventures compared to the ones that you're currently successful in? Do you think that the same principles apply like which your advantages and central features

40:23well, you know, I've actually never been on that big on the exit. both for like my first company Zip2, which basically made internet software that helped the media companies come online. Yeah, you know, and and PayPal, I was the least interested in in selling. and but the the rest of the the board was, so I kind of went along with that.

40:49and but I have no no plans to sort of sell Tesla or SolarCity or SpaceX. In fact, I think it's very important particularly that SpaceX not be sold because then it would lose the the philosophical goals that I have in mind. So, these companies will remain privately going forward or not?

41:11well, Tesla is public and so, in theory, somebody could attempt a hostile takeover of Tesla, although it would be be tricky cuz I own about a third of the company, so they'd have to really I'd have to be doing something really stupid in order to, you know, for somebody to do a hostile takeover.

41:31So, from just doing moderately stupid things, I think they wouldn't be able to take it over. and and then SpaceX I think we'll probably take SpaceX public, but I I'll even going public I'll probably still have a majority share of of SpaceX, so going public would be you know, give us access to low-cost capital, but I wouldn't lose control of the company.

41:56and I think what what with Tesla it's possible at some point in the future, certainly not in the short to medium term, but at some point in the future, it it may make sense or you know, shareholders may wish to sell Tesla. and but I would only agree to such a sale if I I that the fundamental good that Tesla can achieve has had largely been achieved. And that fundamental good is the acceleration to sustainable transport. I mean, that's if one looks back at at Tesla and and says, "Well, what what's the best outcome I could hope for Tesla?" The best outcome I hope is that the presence of Tesla accelerate the transition to electric cars by 5 to 10

42:34years. and and the reason it's important to have that acceleration is that we have a tragedy of the commons problem in in the production of CO2 where you know, we're sort of consuming some CO2 capacity of the oceans and atmosphere, but we're not paying for that at the gas pump. and so you have to you have to bridge that that issue with innovation.

43:01And that's so that's that's the fundamental goal of Tesla and and SolarCity. Yeah. Any women have a question? All 10 of you. Other questions? yellow shirt, right in the middle. Hi. So, I'm currently a student because you know, I can manage energy and space. It's a lot easier for one of us to go to the website than it is to say build a wind farm or auction.

43:36But those are still hard problems. How much time do you have left? Y- Yeah, I think if if somebody's just starting out as kind of a engineer or entrepreneur type of role, then you need to pick something where which is requires low capital. That's why the internet tends to be a good choice.

43:56Alternatively, if you you know, if you could maybe if you can invent something associated with making it a more efficient form of wind energy or solar power or something like that, then I think you may be able to attract venture capital for for that investment. But I I think certainly space and cars would be kind of the equivalent of like graduate entrepreneurship.

44:27But probably not wise to start to to start out in that arena. Just because of the capital requirements are are pretty pretty high. Okay, two more questions and keep moving them up. way in the back. You've got it. Thanks so much for your talk. I wonder if you might briefly touch on a few or even just one thing at a time.

44:50One of the major technical obstacles in creating a reusable spacecraft. And specifically, how is SpaceX going to overcome that obstacle? Is it mainly through new technology, new materials, a combination of technology? How do you do yeah. So, basically the I mean the primary issue with making an orbit class system reusable is that typically even for expendable system

45:20after you've used the best materials and best manufacturing techniques and that kind of thing, you maybe get 2 to 3% of your lift-off mass to orbit. So, if you then start adding in recovery systems and heat shields and all sorts of things, then you're subtracting from that 2 3%.

45:36And and then historically when when there've been there've been many attempts to create a reusable orbital system, they found that after adding in the reusable elements, they get zero to orbit or almost nothing to orbit. So, the real challenge is improving the the engine efficiency, the the mass efficiency, and then just having very lightweight recovery systems so that you can actually end up with let's say at least 2% of your lift off mass coming back at at least 2% of your lift off mass being a payload to orbit after subtracting all of the recovery systems and that kind of thing.

46:21so we do have a design in mind for for Falcon 9 version two that in principle does achieve that. and that's consists of the first stage boosting back to a propulsive landing at the launch site and the second stage having a parafoil to to to a in a landing on either an airbag or kind of like a just some sort of soft soft landing.

46:55and in principle that that achieves the the 2% of payload to orbit. So it'll have roughly 10 times to to the orbit capability lower orbit capability. so that that's what we're going to try to achieve with with version two of Falcon 9. but it's it's it's a very tough thing and and like I said requires optimization across the whole system, you know, engines primary structure avionics recovery system the whole two dimensional it's it's it there's no one silver bullet. You know, it it's a it's an overall system optimization.

47:34Okay, one last question. Put your hands up. Now, if you're over 21, lower your hand. no, that's all right. Sorry. Yeah. I'm a growing generation. it says here that you are the solar city and you can be the future of solar systems and the business. And I've been hearing a bit about what's going on with China and the direction they're going with the solar and the Right.

48:06Right, right. Well, I should clarify with with SolarCity SolarCity cities the nation's largest provider of solar power photovoltaic solar power in the US right now. But it it is it's kind of like analogous to Dell computers. So SolarCity does not make the solar panels itself, but it designs the overall system for a particular rooftop. Does the installation, the wiring, the SolarCity inverter, does the permitting, all that sort of stuff.

48:32So just like Dell computer doesn't make the hard drive or the CPU or that kind of thing. And that actually is the majority of the cost. So if you if you want to install solar power on your home or business, about 40% of the cost is the panel and 60% is everything else. And so that's why I it was important to have a company like SolarCity to optimize the all the sort of thorny bits of the of the what's called the balance of system of installing solar power.

49:02So so SolarCity will buy solar panels from a wide variety of of of manufacturers. But but I should say that the the US companies are actually very strong in the in the in the panel manufacture business. First Solar for example is a US company that's that's phenomenal job on cost of solar power. SunPower provide which is based in Silicon Valley has the highest energy efficiency panels.

49:37So there's there's a huge amount of innovation taking place and most of that innovation is driven from the US. Let me take this opportunity to thank Elon Musk for an absolutely inspirational and wonderful lecture. So please join me in thanking

资源与出处

相关视频

同频道、同主题或相邻年份