访谈·中文精读

马斯克与桑迪·芒罗一对一访谈

Elon Musk Interview: 1-on-1 with Sandy Munro

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

桑迪·芒罗(Sandy Munro)拆过几乎所有在售的电动车。2021 年他开着一台 Model 3 跑了几千英里,在镜头前说了一句很重的话:这套座椅是地球上最好的,没有之一。

同一场对话里,马斯克回忆起早期 Model S 的座椅,说他坐过的车里没有比那更糟的。"如果你想要坐在一个石头做的蘑菇上,那就是早期 Model S 的座椅。"

从石头蘑菇到地表最好,中间隔了差不多十年,也隔着一堆外人看不见的麻烦。

这场一对一访谈发生在 SpaceX 位于德州博卡奇卡的基地。芒罗刚参观完火箭产线,进门时看到一行字,"通往火星的门户"。两个人从座椅聊到一体压铸,从自动驾驶聊到做空机构,把特斯拉当时的工程路线和马斯克本人的判断都过了一遍。

芒罗是那种很少给好评的人。他做了一辈子制造,看车的方式跟普通消费者完全不同,先拆,再量,然后算成本。所以他说"最好的座椅"时,分量不太一样。

产能爬坡期买特斯拉,等于抽盲盒

访谈刚开始,芒罗就抛出一个具体到有点尴尬的问题。

他手上这台 2021 款 Model 3 有些小毛病,而他在路上遇到的另一台白色内饰的 Model 3 完美得挑不出错,"随便谁来做也就这个水平了"。两台车出厂时间只差一个月。

马斯克的回答没有绕:"把生产线拉起来的那段时间,几乎不可能把每个小细节都做对。"

他给了一个很实用的说法。朋友问他什么时候该买特斯拉,他的建议是:要么在最早期买,要么等产能稳定之后买,中间的爬坡阶段别碰。

原因不复杂。产线在垂直爬升的时候,所有参数都在变,而很多参数之间的耦合关系,只有真正跑起来才暴露。

马斯克举了漆面的例子。产线提速之后,原本留给漆面多烘干一两分钟的时间没有了,漆没干透就被推到下一道工序,出问题的概率直接上升。

"如果这些事我们提前知道,就会提前解决。但产线一提速,问题才自己冒出来。"

他也承认芒罗过去那些批评都是准的。Model 3 上一边缝隙对得上、另一边对不上,这种事确实发生过。改善发生在 2020 年底,甚至就在 12 月这一个月里,团队把注意力集中到缝隙和漆面上,质量提升了一大截。

这里有个容易被忽略的背景数字。特斯拉是过去一百年里,第一个实现量产规模的美国新造车公司,上一个做到这件事的是克莱斯勒。

马斯克对这件事的表述很直接:原型车相对容易,而且好玩;把一款可靠的产品以可负担的价格做到量产,痛苦到难以形容。

芒罗接过话头说,他 2018 年拆过一台 Model 3,后来拆 Model Y,那时候有人问该买哪台车,他第一次开口推荐了 Model Y。现在这台 Model 3 让他愿意给出第二个推荐名额。

一个拆车老炮愿意点名推荐,比任何广告都值钱。

座椅为什么必须自己做

芒罗对座椅的评价近乎失控。他说自己开吉普牧马人,在那种座椅上坐一小时就得去找脊椎按摩师,可能还得动手术;而他们在特斯拉座椅上连续坐了几个小时,一点疲劳感都没有。

马斯克解释了设计目标:把压力峰值抹平。

屁股疼的原理很朴素,座椅某个局部顶着你的身体,形成压力峰值,把那一块的血液循环切断,疼就来了。所以座椅设计真正要做的事,是让压力均匀铺开。

他顺手自嘲了早期产品。早期 Model S 的座椅糟到什么程度?他打了个比方,坐上去像坐在石头蘑菇上,唯一的希望是你的身体能自愈,座椅不能。

转机来自设计部门。弗朗茨·冯·霍尔茨豪森(Franz von Holzhausen)的设计工作室做了一把定制座椅,马斯克坐进去,第一反应是"这椅子太棒了"。接下来的问题变成了工程题:怎么让量产座椅拥有手工定制的手感,又不能太贵。

答案是自己做。

芒罗提到,很多主机厂不愿意自己做座椅,理由是应该外包。他的观点相反,凡是人手要接触的部件,都该自己做,那才是知识沉淀的地方。

马斯克同意了,还补了一层原因:跟供应商合作,几乎不可能做出这种效果。

供应商的逻辑是快速出货、摊薄成本。他们会把底下的骨架,那些焊接件或者铸件,反复复用到所有客户身上。结果就是各家座椅外面看起来千差万别,里面几乎一样。你没法要求供应商为你的手感重新开一套骨架,成本不允许。

自己做还有第二个好处,可以持续迭代。今天这套座椅相比早期 Model 3 已经改了很多轮,每一轮都能把上一轮的教训吃进去。

自动驾驶的头号指令:不许撞

芒罗抛出的下一个问题是路。

他试用了 beta 版自动驾驶,发现出问题的地方不在车上:有些路段根本没画线,有些路段线画在正中间,有些地方旧匝道和新匝道叠在一起,锥桶和闪烁的红灯混成一团。他的原话是,如果让他自己开那段路,他都不知道该往哪走。

车子跟着地上的线走,把他吓得不轻。他后来说,那之后他干脆不走最右车道了。

马斯克的回应定了一个原则。

"即使路面标线完全画错,即使一个 UFO 停在马路中间,车也不能撞上去。自动驾驶系统的首要指令是不撞车,这条盖过一切。"(马斯克)

他不否认清晰的标线有帮助,但系统不能依赖标线正确。有人会故意去骗这辆车,车不能被骗过去。安全性必须建立在最坏的路况和最怪的人类行为之上。

马斯克还告诉芒罗,右侧车道偏好是可以改的设置。芒罗当场表示回去就要关掉。

接下来是全场情绪最高的一段。

芒罗体验了当时的 FSD beta。他说自己坐过 F-18 的座舱,也在模拟器里飞过 C-17,知道自动驾驶能干什么不能干什么,但从没见过这种东西。他的评价是:比 Model 3 上现有的版本准得多,风格甚至有点激进,前面有缝它就钻,能自己完成左转,而左转是很多人咬定做不出来的动作。

他给出了一个很大胆的判断:这套系统救下的人命,会比安全带、安全气囊和过去所有汽车安全发明加起来还多。

数字支撑得住这个判断。全球每年大约有 100 万人死于交通事故,约 1000 万人受重伤。这是人类自己制造的最大非自然死亡来源之一。

但马斯克对推送给全部车主这件事异常谨慎。他描述了一个危险的中间状态:系统 99.9% 的时间表现很好,好到让人放松警惕,可它偏偏处理不了剩下的边角情况。人在这种状态下最容易出事。

所以他宁可多测一段时间。把半成品交给用户,比晚一点交付更危险。

代码行数是负资产

芒罗问了一个工程师式的老问题:这套系统有多少行代码?

马斯克的回答让提问者有点意外。他不认为代码行数是衡量好坏的标准,恰恰相反,他倾向于认为行数多是坏事。

他给出的内部记分规则很具体:删掉一行代码加两分,增加一行代码加一分。

在神经网络主导的系统里,这个指标更没意义。真正跑的是海量的矩阵运算和点积,代码行数跟能力之间没有对应关系。他估计整套系统也就几十万行,而且他希望更少。

这个思路跟芒罗的"减少零件"是一回事。芒罗很多年前数过轮罩区域有多少个零件,结论是这里应该是一个整体。两个人在这件事上碰了一下,后面的对话就全绕着这个展开了。

五十个各自正确的答案,拼出一个弗兰肯斯坦

马斯克对轮罩区域那段历史的复盘,是整场访谈里信息密度最高的部分。

问题的起点听起来完全合理。工程师会问:做这一小段车身,最好的材料是什么?做那一小段,最合适的工艺是什么?

特斯拉的材料科学团队可能是全世界最强的,很多人同时在 SpaceX 干活。所以他们给出的每个答案都成立,50 个答案单独看全对,合在一起不成立。

马斯克用了四个字概括这类错误:对错误问题的正确答案。

后果是一连串连锁反应。异种金属接在一起,会产生电偶腐蚀,得加密封;不同材料之间的缝隙得填;有些地方必须用铆钉,有些必须点焊,有些得用结构胶配合点焊。凑在一起看,像一具弗兰肯斯坦。

最痛苦的是填缝。马斯克说,在缝隙之间打密封胶可能是整个工厂里最折磨人的工位。工人干上几个小时,手上稍微抖一下,胶固化出了一点瑕疵,就留下一个小孔,整车密封性出问题,直接变成质量事故。

他用了一个更根本的说法:组织结构的错误,会原封不动地印在产品上。车身被切成那么多小块,是因为组织里就有那么多各自为政的小组。

解法是把它变成一个铸件。

没有缝隙,不用打胶,没有异种金属,车身车间可以大幅缩小。具体数字是:Model Y 的后车身一体铸件,让车身车间面积减少了 30%。Model 3 的车身线上大约有 1000 台机器人,仅后铸件这一项就砍掉了 300 台;等到前车身也改成铸件,还能再砍 300 台。

马斯克特意强调,机器人数量也不是什么值得炫耀的指标,越少越好。

芒罗听到这里情绪上来了。他自己的车间里摆着一套前、中、后三个铸件,摆了 15 年。这些年无数工程师和高管路过,都会说"这个想法不错",然后没有一个人真去干。原因是他们已经把身家押在车身车间上了。

他唯一的不满,是特斯拉没在 Model 3 上就这么干。

马斯克的解释是关于节奏的。Model 3 和 Model Y 是主力产能,是在高速公路上以 80 英里时速跑着的车,你没法边跑边换轮胎。要重做工厂,需要一个不会炸掉公司现金流的窗口。柏林和德州的新工厂就是这个窗口。

新工厂的三件式结构是:后面一个铸件,前面一个铸件,中间是结构电池包。

今天的电芯,是按一袋土豆的方式在运输

结构电池包是马斯克说他从公司创立之初就想做的事,一直拖到 2021 年才动手,因为它是个耦合问题。

在高速研发阶段,工程团队习惯把问题拆开:电池归电池,车身结构归车身。合在一起做,任何一头出问题都会拖垮另一头,等于把整家公司押上去。现在 Model 3 和 Model Y 稳产了,新工厂在建,才有本钱把两件事捏到一起。

马斯克对现状的批评很尖锐。今天所有车里的电芯,都是被当成一袋土豆运着的,它们的结构价值是负的。

负价值有两层。一层是电芯完全不为车身刚度做贡献;另一层是它们必须被隔离起来,防振动、防冲击,否则电芯会撞上电池壳体内壁。

为了这个隔离,你得往车里塞额外的质量。质量本身又拖累了续航和操控。

结构包的做法是把电芯用泡沫粘死,让电芯本身承担剪切载荷。那层粘接泡沫同时扮演两个角色:胶,以及阻燃剂。一个零件干两件事。

为什么能变硬?马斯克用复合材料打比方。上下两张碳纤维蒙皮中间夹一层铝蜂窝,重量没增加多少,弯曲刚度暴涨。电池包的上盖和下盖就是蒙皮,中间密密麻麻的电芯圆柱壳就是蜂窝。

芒罗补充说,他自己拿木块做过演示,在电池日之后放了段视频,解释载荷怎么传递、为什么更安全。那是他们频道播放量最高的一期,因为一堆人跑来告诉他这事行不通。

刚度带来的是驾驶感受。马斯克把几个概念串起来讲:车身的自然频率高、扭转刚度好,开起来就"紧";质量越往中间集中,极性惯性矩越小,转向越利落。

他给了个所有人都能懂的画面:花样滑冰运动员把手臂收起来,转速立刻就上去了。

芒罗的总结比马斯克还狠。他说这不是一箭双雕,是一箭三雕:车身刚度上去了,零件数量下来了,安全性也上去了。

为什么大型铝铸件不能淬火

聊到一体压铸的具体工艺时,马斯克讲了一个很少有人提到的材料难题。

铝铸件想要获得好的力学性能,通常需要热处理,有时还要水淬。问题是,一个一米多长的薄壁铸件一旦经历剧烈淬火,就会像薯片一样翘曲。他用的词是"变成椒盐卷饼"。

老款 Model S 的前后铸件就吃过这个亏。它们必须热处理,于是尺寸没法做大,一大就变形,没法量产。

解法是自研合金,让材料在铸态下就能拿到足够的强度和延展性,跳过所有会让零件变形的后处理。延展性同样重要,因为它决定了碰撞时材料是断裂还是变形吸能。

芒罗在这段里插了一句感慨:他跟人说大型铸件充型只要四到六毫秒,没人信,因为大部分工程师没学过材料。

马斯克接得很快。他在斯坦福读研时本来要去的就是材料科学系,后来把这个计划搁置, фактически 退学了。他说材料科学是工程学科里最有用的一门课之一。

芒罗提到一个对比。福特在 70 年代末到 80 年代初的那轮收缩里,把材料研发砍掉了,从那以后就不再自己发明材料。其他主机厂也基本如此。

他的判断是:大厂现在最该做的事情之一,是把这些能力收回公司内部。电动化之后不再需要发动机厂和变速箱厂那么大的面积,一个小电机加一个单速减速器就够了,腾出来的空间和预算该投到材料上。

12 伏是个退化电压

芒罗问线束。他拆 Model Y 的时候本来期待看到大的变化,结果拆下来的线束看着跟以前差不多。

马斯克没有辩解,直接承认这块还有很多改善空间。

他先纠正了电压问题。12 伏在汽车上是个退化器官。这个标准来自燃油车时代,早就该淘汰。他主张的低压标准是 48 伏左右,理由很实在:以太网供电(PoE)跑的就是 50 伏上下,手机、摄像头这些设备天天用着,没人觉得危险。车身低压系统完全可以到那个水平。

电压提上去,同样功率下电流就小,导线直径可以大幅缩小。铜是钱,重量也是钱。

更重要的目标是把电力和数据放到同一根线上。现在的车还在用大量点对点的专线,未来应该走高速总线,把数据丢到总线上,让节点自己取。线束越少,整车越轻,装配越简单,成本越低。

他也交代了进展。新款 Model S 和 Model X 终于换上了锂离子 12 伏电池,体积更小,容量大得多,日历寿命和循环寿命跟主电池包对齐,"早该这么干了"。

从骑马送股票,到做空机构

话题转到做空,马斯克的语气变了。

他先讲了一个历史。做空这件事的起源,是股票凭证还靠人马在城里之间运送的年代。纽约的经纪商要从芝加哥调一张凭证,路上得走三天,为了不让交易卡住,就有人说"凭证正在马上,我保证它到,你先付钱"。这个约定俗成的漏洞,后来变成了一个产业。

他的问题在于,这个漏洞被用在了谁身上。

"做空今天的使用方式,本质上是针对公众的。绝大多数股东根本不知道做空存在,知道的人里又只有极小一部分懂得怎么用空头头寸获利。这实际上是打着市场效率旗号的对公众的攻击。我个人认为它不道德。"(马斯克)

他提到一个反差:全世界九成以上的公司是私营公司,私营公司没法被做空,可照样运转得很好。所以那套"做空提升市场效率"的说法站不住脚。恶习披上了美德的外衣。

特斯拉被这套机制打过两次。一次是 2013 年,一次是 2017 年到 2019 年。他描述的做法叫"做空加抹黑":先建立空头头寸,然后动用一切手段把公司名声搞臭。他说那段经历的强度大到让人对人类失去信心。

芒罗在这件事上的态度更直接:一群人可以仅仅为了完成本月的业绩指标,就把一家公司碾死,他从来不是做空者的粉丝。

领导力不该从商学院空降

芒罗主动提起了马斯克那句关于 MBA 的著名评论,说自己多年来一直是同样的立场:员工想读硕士,公司出钱;想学机械,出钱;想学木雕,也出钱;想读 MBA,一分不给。

马斯克把逻辑讲完整了。

问题不在商学院教的东西,在于一条被走偏的路径。通往领导岗位的路,应该是一路做有用的事爬上来,而不是读完一个高知名度 MBA 之后空降成老板。

空降的人往往擅长做幻灯片,讲得很好,但不清楚东西到底是怎么做出来的,也不知道做好产品真正需要什么。

他补了一句限定,说自己没有全盘否定 MBA。他在宾大读的是沃顿本科加物理双学位,对商学院有直接接触,还给本科生和 MBA 当过两个学期助教、批改过作业。

他真正反感的,是那种"我要直接空降当老板,而不是挣来这个位置"的心态。

电动车超越燃油车的时间点

芒罗提到通用汽车宣布 2035 年全面电动化,然后问马斯克怎么看拐点。

马斯克说,他几年前就公开讲过,纯电对燃油的交叉点在 2030 年。当时台下不少高管听完跑来跟他交流,而华尔街分析师的预测是 2045 年电动车占比 15% 左右。

现实跑得比分析师快。加州当时的电动车份额已经到 8% 并且在往上走,欧洲开始在政策上直接排斥燃油车。他的修正判断是:十年之内,多数新车销售会是纯电。

他顺带说了一句不太客气的话:很多人希望一切维持原样,回到所谓的"过去好时光"。这个愿望跟物理规律没关系。

没有自动驾驶的车,会像马一样

访谈快结束时,马斯克给了一个远景判断。

他说,未来一辆没有自动驾驶能力的车,受欢迎程度大概会跟马差不多。马还在,人们还养马,但没人拿马做日常通勤。

芒罗接了个很有意思的梗。他说从前有人喝醉了在马车里睡着,马会自己把他拉回家。从这个角度看,自动驾驶只是把这件事还回来。

马斯克顺着往下讲,说这两件事他们真的打算做进车里:如果检测到你睡着了,车会直接送你回家;如果检测到你的身体出问题,比如心脏不适,车会直接开去急诊室。

这一段听起来像玩笑,但背后的数字一点都不轻松。全球每年约 100 万人死于车祸,约 1000 万人重伤。马斯克重复了那句话:得赶紧干。

芒罗最后说,这辈子最让他兴奋的事,是坐在那台车里看风景。以前他只能盯着车道线,现在可以抬头看看周围,手还放在方向盘上,随时能接管。

这句话其实是整场访谈最好的注脚。技术成熟的标志,不是让人惊叹,是让人放松。

这场对话能拿来用的东西

  • 打算买新势力品牌的车,避开产能爬坡期。要么订最早那批,要么等产线稳定半年以上再下单。中间那几个月,装配一致性最差。
  • 评估一家制造企业,别看它有多少台机器人、多少行代码、多少个零件。看它敢不敢把这些数字往下砍。删掉一行代码加两分,增加一行加一分,这个记分规则可以直接搬到自己团队。
  • 做产品决策时警惕"对错误问题的正确答案"。每个子系统单独看都最优,合起来是灾难。判断信号很简单:接口处需要大量补丁、密封、转接,就说明切分错了。
  • 凡是用户身体要接触的部件,尽量自己做。外包给别人,底层平台会被复用到所有客户身上,你永远拿不到那个手感。
  • 遇到自动化系统的边界问题,先看基础设施。路面标线画得乱七八糟,再强的感知也救不回来。系统设计要以最坏输入为前提,不能以环境配合为前提。
  • 别在半成品阶段大规模推给用户。99.9% 的可靠性最危险,它让人放松警惕,然后被剩下那 0.1% 干掉。
  • 带团队别迷信空降的商业教育背景。让做过实事的人上来,产品里印着的,永远是组织结构的形状。
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)。

0:04hey boys and girls we got a special treat for you today i've gotten the opportunity to to meet with mr elon musk and today we're going to talk a little bit about his aspirations we luckily got a tour through the program here at and brownsville and i i just seen it says bocachia a gateway to mars so it's it's kind of it's kind of really exciting i was really really happy that i got the chance to to see the the the product that's being manufactured here these rockets it's spectacular the way they're doing it and it's totally different than anything i've seen with rocket production at any other company so with that i'd like to introduce mr musk and

0:50and i guess we'll maybe you could answer a few questions or actually i'd rather i'd rather hear from you than than talk i i've been doing a lot of talking lately and like i've had criticisms in the past of of you know were you talking to an interview or something yeah no i wasn't in an interview was when i did my reports and and the and it was what it was okay and even on the the model 3 we have i wanted to show you the gaps on one side but not on the other i i don't understand how that can happen so but i thought your criticisms were were accurate yeah well and so i've got a question we've been traveling all over the place and suddenly we're fighting we've got a

1:37sticker on the side of the car that says the mineral life and people are coming up and showing us their car and telling us about their experiences with a tesla and why not yeah these guys are really deep into into loving this thing and it doesn't matter whether it's that eight-year-old or six-year-old kid what that came up mister are you sandy monroe people know and and they they're deeply involved in the tesla and yet for some reason or other we have a couple of little problems with our car and then i see another car and it's a perfect build if i would have had that any any kind of comment or response or whatever from that would be greatly appreciated

2:14yeah well it it took us a while to kind of iron out the production process especially during a production ramp it's that's like people like friends cannot ask me like when should they buy a tesla like well either buy it right at the beginning or when the production reaches steady state but during that like production ramp that's it's super hard to you know be like in vertical climb mode and get everything right on on the little details it's just a super difficult thing so if you really want like things to be dialed it's like actually very very early cars or once once production is leveled off that's when it's going to be best well we noticed that we have a we have a 21

2:54at the model 321 and ours had the little problems but but but at the end of the day this guy's car was fabulous paint jar was spectacular sure but the paint was it was a big deal but now this other car i saw and it was a model 3 with white interior which i personally think is the better color than we got black but that that thing was absolutely pristine perfect that's as good as anybody anybody could possibly do i just don't understand it mine was built this month his was a month later mine had problems his is perfect yeah we actually did improve gap and paint quality quite a bit towards the end of last year even even in the course of december

3:38there was like yeah it we were able to really focus on it and improve it a great deal like one of the things that was happening when we were like ran from production was the paint wasn't necessarily drying enough so it's like well if you go faster you know it's it's like you you just discover these things like so if we knew them in advance we'd fix them in advance but but like like you ramped the ramp the line and and then the paint that had had an extra sort of minute to dry or two minutes or whatever now it doesn't have the two minutes and so it was more prone to you know have two issues that might be correct correct this is like one example

4:21yeah and yeah it's it's a production as hell like the like the real the thing that's i think unique for cars for tesla is achieving volume production it was like of any american startup car company i i think tesla's like the first to achieve volume production in a hundred years basically like i think chrysler was the last one yeah and so like prototypes i've said it's probably like prototypes are relatively speaking easy and they're also like fun like prototypes are easy and fun and and then reaching volume production for with a reliable product at an affordable price is is like excruciatingly difficult yeah so you've done a good job i mean i've

5:08seen the difference between the one we had in 2018 then the model y which was that's the first i i never recommended anybody else's but the one the model y when we tore that down people were coming up hey which one should i buy and i said i'm not into doing that but if i'm going to buy an electric car and we've seen everybody's i mean we've torn every everybody's electric car apart if i was going to tell somebody to buy one and he had had to have one i would suggest the model y and lots of people did that this one this model three we've got this may be the second one that i'm going to suggest you know you can either have one of these two and then i won't have to worry

5:50about you coming back and telling me there's something wrong with my skull or something that they they both they both work really well yeah and i have to tell you we've been we've gone about almost 6 000 whatever i can't remember 6 000 plus miles so far in a couple of days in a few days and the seats in your car are phenomenal absolutely phenomenal i i can't say enough about them i i i drive a jeep wrangler yeah and and if i had to sit in that seat for an hour i'd have to get a chiropractor and and and probably surgery these seats we were sitting in them for hours and hours and hours and hours and there's no fatigue they're brilliant and you make them yourself yeah yeah

6:39absolutely so i mean what we're really trying to do there with the seat is and we put a lot of effort into this is minimize any peaks and any pressure peaks so it like evens out the pressure yeah like if if your butt hurts it's basically going to be because there's there's some part of the seat that is producing a a pressure peak yeah and that's that's gonna just cut off your circulation and make your butt hurt but but there's there's been quite a few iterations in fact i mean we've the early model s's had i think probably the worst seat of any car i've ever said

7:20it's like if you wanted to sit on a snow and toadstool that was the early model s seats now if you sat on it for an ever it barely i guess it's better than a stone toilet stove so because if you started long enough eventually your body your heart would win eventually yeah because your butt can repair itself on the seat cannot yeah so eventually it would be like acceptably comfortable but it would be a very difficult wearing process and so we you know we it's like we try to go from okay from stone toed stool to something that is just i don't know lap election it just feels great you know and yeah that you can sit in for a long time time and it doesn't it doesn't

8:00it's still comfortable that that was a long journey and a lot of effort yeah well yeah it paid off i like i said we do a lot of work on different things and i would say without a question of a doubt the seats i never read we did no room in the back seats but we had it filled up with junk but but i would suggest that that seat is in my estimation for my body and when it is the best seat on the planet there's nothing better than that and a lot of the oems they don't they don't want to make seats because they say that's something that should be outsourced and i believe that anything you touch anything that that you're going to interface with yes

8:43has to be has to be made in-house because that's your profound knowledge so yeah absolutely and you can iterate on it and keep making it better and this this this particular seat it's it has improved quite a lot from the early production model three and one you know early pressure model three two to present day it's improved a fair bit and like one of one of the breakthroughs was like actually franz when hull thousand had a seat that the studio had done and and i sat in that seat i'm like wow this seat's great we're like okay let's figure out how to make how do we get a production seat to feel like a bespoke handmade handy tailored seat without costing a ton of money

9:22so you just can't it's very difficult to do that if you're working with a supplier you know so well near impossible because they got to get it cranked out and moving and quickly and then what they want to do is they want to take the underpinnings those weldments or castings if they haven't gone that far they want to use those over and over and over again for everybody and so consequently everybody may look different on the outside but inside probably not so much so that's that's good stuff but i have to tell you about the best stuff and okay so we had we had the beta autopilot system and and we did have problems but the problems weren't with your

10:04system they were with the roads people are painting either they don't paint in certain areas or they paint right in the center so how are we going to get legislation to make it so that things are consistent between states in the old days it probably didn't make much difference but now we're moving into self-driving and well i mean i think for self-driving even if the road is painted completely completely wrong and a ufo lands in the middle of the road this the car still cannot crash and still needs to do the right thing so the like what really matters like the prime directive for the autopilot system is don't crash like that that really overrides everything so no matter

10:53what lines say or how the road is done the the thing that needs to happen is minimizing the probability of impact while getting you to your destination conveniently and comfortably but but the prime directive absolute priority is minimize probability of of injury to yourself or to anyone on the road on the pedestrians or anything like that so and it's really it can't be dependent on the road markings being correct or anything like that it's just going to be no matter what it's not going to crash well i still think that that the road commissions in one eye better shape up because like i said corey and i had an issue and luckily and luckily i got out of it

11:37but what happened was they painted all kinds of lines all over the place there was an old off-ramp and a new off-ramp and a bunch of cones and some flashing red lights it was totally i have no idea if i had to take that thing and knew i had to take it i would be totally confused as to what to do where but when the car came into it the the lions are telling it okay go here and that scared the living daylights out of me if we would have crashed if it would have been somebody i used to race and i kind of like can get out of stuff and and even though i'm old i'm kind of agile and but if we would have crashed the press would have instantly said it was

12:20all your fault and in essence the fault lies in this massive mess that that we were trying to keep up with and then quite frankly i stopped driving in the in the right-hand lane that was it i mean the car wants you to keep going back there because it wants you to follow the rules but you can change that it's a setting oh really yeah you'll have to show us that because i don't like i don't like that one but but it's it's it's following it's following the rules and that's great but you have to have the roads at the same sort of thing so i want to make sure that we've got this on tape or on whatever not tape but but video because i'm guaranteeing you

12:58somebody is going to screw up somewhere and and and people have to know it's the road that can bugger the things up as well as whether your stu your your autopilot thing i mean it would certainly be helpful to have roads with accurate markings and everything but like i said the really for self-driving it's got to be even if somebody tries to trick the car they do not succeed in tricking the car right because you know people will do weird things so it's got to be maintain safety no matter what and don't let yourself get tricked that's yeah so well maybe that's the best segue ever because what i want to do is i wanted to talk about i i actually john

13:38what's john's last name no no the guy that led us gave us a ride yeah okay well i'll just say john l.a silicon valley yeah la silicon valley whatever i thought anyways so anyways he he gave us a ride okay and and we taped it and you can hear me going you know my voice went up about three octaves i was so excited corey said he's got a sore ear for me giggling or whatever i have nev i've sat in in f-18s i didn't fly them but i was sat in them and i saw how everything was supposed to work on the ground yeah i i flew or through a simulator i flew a c-17 i know what you can do and what you can't do what autopilot will go to and i never seen anything never ever

14:31seen anything quite like what you've got in the new self-driving thing this this is just absolutely brilliant this should get into the marketplace as fast as possible it's it's it's accurate it's much more accurate than what we have in the in the model 3. it's accurate it's it's kind of aggressive because if there's a hole it'll it'll find that hole it makes left-hand turns which i've heard from everybody can't be done i mean these are the things that this is the this will save more lives and airbags seat belts and anything else that anybody's ever ever gotten because that's correct yeah i i really i i was so impressed i i couldn't believe it and

15:09i have some videotape and i just asked you for using it i i want to put that out i want to i want the rest of the world to know what the new standard is because in the auto world some people are going to win some people are going to lose and some are going to fade away the more they fade away the worse it's going to be for the for the general population but i'm telling you that that system i don't know who you developed or how it developed but that is absolutely developed like we developed the hardware and the software we've just got we've got a very talented team that we bought from scratch at tesla for autopilot software and autopilot hardware

15:47and yeah it's just we've got really a lot of talented people again i gotta ask a technical question how many lines of code is in that in that thing a bazillion a trillion whatever what well i don't think lines of code is necessarily a metric of goodness well maybe but i mean it's so it's so it's so intuitive either that or it's ai and it's cognitive knowledge or something i don't know but yeah i think generally the i would consider lines of code to be

16:21i like a large number of lines of code to be bad not good and in fact i would generally give like two points for deleting a line of code one point for adding a line of code so and and then like the lines code thing when you when you have neural nets where you've right i got it it's it's not you know it there's a whole whole lot of statistics going on basically a lot of dot products yeah going on the lines of code is becomes especially not a merit figure merit for when you've got a lot of complex neural nets operating you have a lot of matrix math i mean that said i think we haven't actually counted it but i mean it's maybe i don't know a few hundred

17:05thousand lines of code that's amazing i'd like it to be less you know well your your philosophy about getting rid of lines of code i have a philosophy about getting rid of parts so the more parts you get written in fact a long time ago when i counted up the number of things in the in the wheel area i said this should all be one part and then totally agree yeah and you yeah you got it actually i can tell you how that that problem arose and it it it was you know

17:35like first of all i think you can generally see the the errors the organizational structure errors they manifest themselves in the product so for the you know the the sort of wheelhouse areas of the body there was a lot of engineering done and there were a lot of right answers to the wrong question so somebody would say like well what's the best material to make this little section of the body out of and what's the right materials make this little section and and i think we've got probably the best material science team in the world at tesla and then actually a lot of them also just do work at spacex as well yeah i know i'm coming to that question

18:18too or yeah yeah but so the the engineers would ask what's the best material for this this purpose best behavior for that and and they got like 50 different answers and they were all true individually but they were not true collectively right and so and when you when you try to join all these dissimilar metal you know similar alloys and you you have galvanic corrosion so you cannot you've got to have better seal and you've got gaps that you've got to seal and you know i've got to join these things and some of them don't some of them need to be joined with rivets some need to be joined with spot welds something needs to be done with resin or resin and spot welds

18:54and it looks you know frankly it looks like a sort of bit of a frankenstein situation when you when you look at it all together and and it's i can i can't emphasize enough the nightmare of of ceiling in between the gaps that is like that might be the most painful job in the whole factory is is spackling on the the sealant yeah and then you make you have a little if there's like if there's like a little error with curing the sealant or like like just somebody who's been working you know for several hours makes a slight mistake then you've got like it now you've got an mbh issue because you've got a little hole and you've got a leak issue and it's it's just like this is terrible

19:34so i mean you can muscle through it and we have but it's just way better to have a single piece casting and then you don't have any gaps no sealant you don't have dissimilar metals and you can reduce the size of the body shop dramatically right so just just having the the rear body castings for model y allowed us to reduce the the body shop by 30 percent yeah so there's roughly a thousand robots on the model 3 body line which by the way is also not a figure of merit you want fewer things not more yeah we got rid of 300 robots just with that rear body casting yeah and then when we go to the front body casting we'll get rid of another 300 robots

20:19right so i don't know if you well you probably don't have time for my stuff but i have in my shop for 15 years a rear casting a center casting and a front casting to make a car yeah i always totally agree thousands of engineers and i mean big big time executives walk by it hey have you ever thought of doing oh you know sandy and they don't do it and they don't do it because we got this body shop sure and and because they've continuously hung their hat on this thing they figured you know this is the way it should go but when you came up oh i will tell you i'm very disappointed i thought i was going to see a single piece casting in a model 3 as

21:03well i thought you were going to shoot the two and then glue it together but that's all right i mean at some point i'm getting over probably switched to a single piece casting but it's it's like we i think we need to get the probably the texas factory and the berlin factory going yeah and like we just we need a like we do have an issue of like it's hard to change the wheels on the bus where it's going to 80 miles an hour down the highway yeah so you know model 3 is like most of our or what is most about volume model y will become it will exceed model model three but it's just we we just need basically an opportunity to kind of redo the factory without blowing

21:49up the cash flow of the company yeah i mean i understand the you've got a lot of stuff in place so we're looking forward to seeing what's going to happen in berlin that that's one i'm really really interested in because quite frankly you're a lot closer to you know you're a lot closer to the the the casting companies and germany has fabulous tool maker so i'm expecting to see lots of good stuff out of that are you going to have to or maybe you don't want to talk about that but are you going to have three castings then or just a two well effectively there'll be a rear casting a front casting and then the center will be a structural pack yeah so structural pack is also this is

22:32an important thing that i've been wanting to do for since the beginning of the company but it was it was always tough to kind of like because it's a coupled problem so like and and when when you're in sort of high-speed r d you're it's you really you often want to decouple the problems can try to solve electrical issues in the battery and you're trying to solve structural issues and if you bring them together then it's it it's hard you know you're you're putting the whole the whole company at risk basically yeah but now since we have model three model y in production we and and we have a new factory that's getting built this is our opportunity to say okay now

23:10we're going to do a couple problem where we're going to combine the the battery pack and the basically body chassis the primary structure the primary structure and we're gonna transfer share through the the the cans of the of the cells and just to you know shear transfer which is gives you a super stiff you know real great moment of inertia it's like a carbon they do this with carbon fiber you know and composite sandwich like have an aluminum honeycomb with upper lower carbon face sheets and you get incredible sheer transfer unless you're a great moment of inertia for it you get a real stiff big plate and so getting dual use of the the cells and like so the cells become

23:55structure and that's like what a point i was trying to make at the a battery drape presentation but i think a lot of people didn't quite understand why that's such an important thing you know the the cells today in every car are carried like a sack of potatoes they have they actually have negative structural value because not only they need to be not only they don't serve to aid in the structure of the car and they they have to be isolated from the rest you know from the rest of the car so it's isolated from you know vibration and shock loads that kind of thing so so then you've got to put mass into isolating the cells because otherwise they'll bang against

24:34the side of the the battery casing and that's that's not good but by essentially binding the cells in there and where the bonding foam serves as both a an adhesive and a fire retardant yeah it's you you basically get you get you know two bows with one stone exactly yeah so multifunctional designs is something i really really try and get my the companies that i work with i try and get them to do that and we got a ton of people right after the the battery your battery show saying oh you can't do this you can't do that so we made one up or made a wood blocks and went down and said this is what's going to be and then we i talked a little bit about you

25:20know what the effect is and the loading and transfer of loading and how much safer and stronger this is going to be on and on it's the highest rating i think it's the highest rated show we've ever put out i'm just explaining you know why this is such a good idea because you got you got people you know designing cars all over the planet and they're telling me why it's not a good idea and i'm saying tell me again i mean really you're an engineer that you this is the kind of stuff that that really needs to have happened yeah i think it likes to be helpful to show and tell because if you say like you know shear transfer using a steel cancer shear transfer

25:59between upper and lower face sheets what does that actually mean like well let's show you you know this is what happens if you've got just a little floppy sheet or or you just have like a limited number of stringers for shear transfer you can see it's like it's still it's quite bendy but as soon as you have a whole bunch of cans or honeycomb or anything like that and you bond an upper lower face sheet it gets crazy stiff like you can't bend like you just put them together like one thing's floppy and the other thing is stiff as hell and that's really what you want and then this this will give also give like like the torsional rigidity won't be

26:35much better and so if you're like you're trying to improve like the the the ride and feel of the car you say like you know what what's the what's the frequency of this car is it like a like a sort of real bendy spring or is it like tight and that feeling of like is this car tight really if people don't understand why does this car feel better than the other cars like well because you've got your natural frequency is high and you've got good torsional rigidity and then another factor is like what is your problem moment of inertia so like basically to what degree is like the mass group towards the center as opposed to spread out towards the outside

27:11and like you can see this like with like an ice skater where if the ice skater holds arms out it you rotate slowly brings arms in rotate fast that's what polar that's that's what's meant by polar moment of inertia exactly so it's like can you rotate this car fast you can if you bring the mass to center right and and quite frankly we've had lots of people discussing you know the pros and cons here but to me you were saying well we got two for one but you really got three for one yeah because now you've got body stiffness that's gone up and all the other stuff that's going along with it right and and then on top of that you've got i i when people say well what's it like

27:51to drive the car because again we've stopped in lots of different places and i tell them if you if you get a 7 series bmw and drive that that's what you're basically getting it's the same sort of a suspension feel as that and it and you don't have to you're not muscling anything around everything is tight and perfect and tight's a good and then if you when you go to the when you go to those three let's say three structural members that i mean if you put a body on top of that it's going to be a perfect build and as a as a guy who had to work in body shops and as a guy who used to make molds and dyes and whatnot it's really difficult to take a whole bunch of

28:33pieces and glue them together and there's no nothing right i'm just using fixtures and then some you're gonna have to stack up i mean that's the issue yeah it's like if you've got a whole bunch of separate parts and each of them has got a given tolerance even if that tolerance is tight and it could be like you know 0.2 millimeter tolerance but you've got like 50 parts okay now yeah you know and you multiply them together yeah and you can't win exactly you're gonna also have like just variants between cars so that's for sure like one of the reasons why it's best to just combine the parts and not have separate parts and then also just go for extreme

29:06levels of precision like you know one of the examples you know we use tesla is like lego so like lego is super precise because the press fit i think they're precise down to you know about a quarter millimeter or less you know basically about ten thou and and each one is exactly the same yeah it's going to be like if the lego if lego's like two presses is is too soft or too hard like it was too soft the precipitate won't work it's too hard you can't get it on right so but they can get make something to that that is like a tiny fraction of a millimeter accurate and it's a low-cost toy right you know by 20 bucks for a lego set and and so so like if lego can be

29:50that precise and so could a car exactly yeah actually i i got to tell you i i applaud bmw for putting out the i3 but it ugly but the but the thing that we got from it was that carbon fiber body was absolutely perfect and when you have if you had the stiffness of the castings down below and you take a carbon fiber basically structure and put it over the top absolutely every door is always going to be perfect you won't have to have fixtures you won't have to have anything it just goes boom done that's ultimately what i think a perfect build is probably going to go for but most people don't want to do it because they're talking about oh carbon fire is so

30:36expensive no it's not it's not really that expensive and and at the end of the day you have to take it and look at it a total account it cost yeah how much does it cost to basically where's all the tooling and whatnot how about all the scrap that you get on and on and on yeah i mean there are some other issues with carbon fiber which is that you know carb fiber with resin has a coefficient of thermal expansion which is basically zero and if you have parts that are say aluminum or steel that's you know aluminum's got quite a high you know cte stainless has it's like about halfway yes you know so it's like so then you know if you have like a hot

31:16versus a cold environment like you can't have your doors jam yeah so having you know matching matching cte is pretty important for things like a door frame all right you know so that's you know if so if we're using say steel we probably need to stay with a steel door because otherwise we're going to get like cte issues yeah but at the end of the day we found out the tricks that that bmw used so they've got cross car beams which is going to shrink the most and they're embedded right into the carbon fibers how is that possible this will rip free we tried this when i worked on the boeing 787 and stuff yeah and we tried that and when we you know you do the thermal expansion

31:57and contraction it rips free from everything but clever design or clever engineering on bmw's part made it so that you could do that so i'm not saying you know run out and do this right now but eventually you guys are material science experts i mean you make your own we know sure a little bit about what's inside those electric motors and and actually the the aluminum used for the for the mega castings these things all nobody has anything like that nobody's using it most people i mean they develop our own allies like so it's like some of the challenges with with doing like a large aluminum casting there is you don't want to have a a a sort of

32:36a heat crate or quench step afterwards so like like often in order to get good properties from aluminum you need to you know yeah yeah well if yeah if you want to if you you want to heat heat treat it and and then the heat treating and any kind of like arrow certainly no water quench or it's it's going to cause the the

32:58things get a potato chip on you it's going to basically warp so we had to develop a a custom alloy to make sure we could get good material properties but not require a any step after the casting that could distort the shape that was actually that's an important that's like a very important thing for example like the model s and nx castings they for the you know original model s not the new one but the for the original they required a heat treat step right afterwards and and so it was always a pain in the neck because we couldn't really expand the casting because it would it would run pretzel on us like just potato chip or pretzel it's like not

33:37not you can't scale that so yeah threatened about development alley that was okay and had good also had good elongation properties in the event of a crash there's like kind of a hard materials problem so well it worked out really well and the no one believes me i know a fair amount about you know casting and why not and no one believes when i say you know it fills up in four to six milliseconds oh that's impossible really well okay fine that that kind of stuff is lost on most engineers because they don't take materials so i wanted to be a mental urges yeah mature science is i think one of the most useful classes you can take in engineering for sure absolutely

34:16i mean the the area that our department i was going to be in at stanford for grad studies before i kind of put that on hold slash dropped out basically was material science yeah so yeah well i'm a big fan of material science well i i am as well i think that i think that that's something that i saw disappear when i was at ford motor company because of the things that were going on in the late 80s and sorry the late 70s and early 80s and they just had to shed some stuff and that went away like they don't they don't invent materials there anymore and actually no one else did either or does either so i think that those kinds of things are the things that maybe the bigger

35:03oems should be thinking about i think bringing stuff back in especially now i mean i don't need the size of an engine plant and a transmission plant and i got a little dinky electric motor and a and basically a two-speed gearbox or one speed gearbox i mean that's that's that's that's going to be a big big difference i don't know what they're going to do with all that stuff i really don't know how how they're going to start implementing this i mean you may have heard that general motors said that 2035 they're going to be nothing but electric cars that's cool well that's that's a pretty big step i've been saying for a long time in fact i had a lot of inch a lot of executives

35:48come in after i gave a speech and said this was a couple of years ago now and i said that the crossover point for electric over ice is 20 30. and basically that was because of tearing apart your car the model three majority of your car would say crossover point how you're finding it fifty percent of the vehicles are going to be are sold are going to be or more than 50 percent are going to be a pure eevee or they're going to be hybrid and and hybrid like you know for sure i think it probably ten years it's probably a majority ev yeah i i truly believe that but when i brought it out i mean the

36:31analysts i don't want to get too far deep into that but analysts were saying you know maybe somewhere around 20 45 there'll be about you know 15 percent ev's and whatever and now we're looking at california with eight percent right now and going higher and then you've got europe saying forget it you can't drive here if you've got a if you've got a nice vehicle on and on and on so a lot of people were hoping that everything was going to stay the way it was or basically go back to the good old days as it were i was not i was i i was in your camp like i say when i got your car initially and there really been i didn't care one way or another car's a

37:14car we take it apart we look inside we find out how much it costs and what the techniques and technologies are and then we must move on sure with yours initially i don't like this i don't like that then i drove it because i was the last one to drive that vehicle and he said you should really take this thing for a ride and see what it and i took it into a parking lot there's a college that's near us and i zipped around and whatnot and i'm i feel young again this i didn't expect that that that excitement and then we started looking at the electronics and how much better even the wiring is in well i think the wiring was so we have a lot to improve on the wiring

37:55front i think there's a lot more that can be done to improve wiring yeah well right now also like 12 volts i mean what are we still doing with 12 volts 42 is where everybody should be but i think 48. more 48 yeah but but at the end of the day we we we need to do something that'll that'll reduce the wire diameters and stuff like that and yeah and that's just nothing but cash in the bank yeah so so i'm a big fan of that as well with the new sx we're also we're we're finally transitioning to a lithium-ion 12-volt so oh good yes it's an excellent idea smaller box it's got you know way more capacity and and the the the the calendar and cycle life

38:34match that of the main pack so we should have done it before now but it's great that we're doing it now also like this is one of those like you know inside baseball victories that's yeah kind of a big deal but yeah like the 12 volts is like very much a vestigial voltage yeah and it's like absurdly low so i mean even like basically powered ethernet is like around 50 volts yeah and so so you can have like powered ethernet nobody's like sweating at that like there's a power to ethernet this like phone here nobody's worried about like powered ethernet like whatever 40 48 50 volts that's that's really what the car the car's low voltage system should be at

39:16yeah absolutely i agree in fact when we talk about the y i was expecting to see that because you'd said that you're getting rid of weight and and the length of wires and whatnot and when i pulled it off the harness looked kind of similar so it is what it is you really want to you really want to put you want to put power and data over the same wires that's right and and have it be a high speed like higher speed than can bus right and so so you can like basically dump data on the bus instead of having all these point-to-point wires absolutely that's that's kind of like what i was kind of hoping i would see when we we took apart the y but again you say

39:50same thing it's hard to change the tires when you're going 80 miles an hour so i guess the the last thing is probably probably the most controversial you made comments which i've been saying for years at monroe if you if you want to if you want to get your masters i'll pay for it if you want to take a course in mechanics or something i'll pay for it if you even if you want to take a wood carving thing i'll pay for it but i will not pay for an mba i won't do it ever never so you made a comment and i i started tap dancing because people listen to you i don't get the same kind of respect but people listen to you and i really really was excited when you when you

40:35talked about about that so if you can elaborate on that a little bit that that would be great yeah i think just generally the the path to leadership should not be through you know basically nba business school situation it's it's like it should be kind of work your way up do useful things and you know and there's a bit too much of the somebody goes to a high-profile mba school and then kind of parachutes in as the as the leader but they don't actually know how things work they you know they could be good at say powerpoint presentations or something like that and they can present well but they don't actually know how things work because they do not you know they're

41:28they're kind of like just not aware of what's really needed for you know to make to make great products so i mean i don't want to trash mbas too much here and i actually do have a dual undergrad wharton undergrad and physics at upenn so you know i have like you know it's like i have direct exposure to to you know to business school and i went you know to do undergrad business school with with physics but and i was a teaching assistant for two semesters and i graded mbas and undergrads so but i think it's just a little bit too much like like people look at mba school as like i want to parachute into being the boss instead of earning it and like i don't think that's that's

42:16good like yeah well speaking of not good and this might be the last thing i've never really been a fan of short sellers i i don't know whether you have a comment on that well i think that the you know there's there's very few areas in life where you can you can sell things that you don't own and the short selling really really where you can sell shares that you don't own when i said it was vestigial i mean it came from an era when stock stocks were traded by people traveling on horseback to exchange stock certificates and in order to have the the transaction speed not be like not take weeks somebody would say well the the stock certificate is coming

43:01on that horse i don't have the stock certificate right now but i promise you that i'm getting the stock certificate and the writer is going to be here i mean new york writers going to be here from chicago in three days and then and then i'll be able to give you the stock certificate that's where this whole silly thing arose but then but then the problem is like the way short selling is used today is it's it's kind of like a it's frankly used against the public you know and so the most people aren't aware that short selling even exists and then the ones that are very few of them know how no actually how to use it it's basically like.01 of stockholders know how to use

43:39short positions to to get ahead and it's so it's it's like i think effectively attacks on the public i think it's immoral myself i just don't think it's right at all yeah a big group can come in and crush a company simply because well we're going to make a lot of money i haven't made my haven't made my quota this month or what have you so i i really don't care for it yeah i mean tesla was under a massive attack by the short and distort where they take a short position and then they do everything possible to trash the company in every six ways to sunday and they were successful and and this has now happened to tesla twice it happened in 2013 and it happened in in basically 2017

44:21through through 2019 and i mean the intensity of the attack was was was crazy and i was like man this this you know it would like cause you to lose faith in humanity the the greed of this wish this went on and like we don't have shorting in like private companies the vast majority of companies over 90 are are private and you cannot short them and yet somehow we get you know private companies get things done so there's there's a a a a perniciously false effective markets argument made for a shorting and it's it but it is vice disguised as virtue well at the end of the day we had a common a common adversary and i'm really hoping that now because he's bad on oil and

45:16and biden is basically clobbered oil in a lot of different directions i'm sure that he'll be skeptical about doing any more trades on a short end anyways i can't thank you enough i i'd really like to shake your hand thank you so much thank you for for giving us the time and i i can't say enough about a lot of stuff but i really can't say them i the most exciting thing that's happened to me in a long time was driving that the the self-driving the self-driving i'm telling you it's just sure it's it blows your mind are you gonna are you gonna let anybody else have that yeah of course i mean we're gonna roll out the pulsar driving to the whole fleet

46:00and make it available to the whole fleet we're just being very careful about the testing and i think there's actually sort of a dangerous middle ground that would be careful of where the the system is is good like 99.9 of the time and and that could lead people to be complacent and and then but then that one time where it's got issues that that you know we don't want people to basically it can be so good that you you get comfortable but not not initially good enough to handle all of the corner cases so we we want to just make sure that in that transition to full self driving it's really we're taking as much care as possible yeah well i'm i'm i was happy with what

46:43we had on the on the autopilot it's the first time i've ever gotten a chance to see what the scenery looks like i always just do lines now i can have a look around and feel comfortable or confident that my hands are on the wheel and if something goes wrong i can correct but i can also enjoy what's going on yeah i mean the like you know so publicly but basically the a car that does not have self-driving in the future will be about as popular as a horse horses are they're still horses people have horses and they're they're but they're they're it's a horse is not what you use for day-to-day transport no yeah but even even a horse will get you home sometimes if

47:26in the olden days they'd get drunk fall asleep in a buck board and no horse would take him home self-driving as far as i'm concerned we're just getting back well i think actually that's that's one of the things we're we're going to program into the car that if if you if you if you fall asleep in the car it'll just take you home yeah like it's most likely that's where you want to go yes that's a great idea yeah and then we're even gonna have it like so if it detects that you're having issues it'll take you to the hospital like you say like you know just like yeah having a heart attack or something yes it'll just take you straight to emergency room well i think

48:00i mentioned this already i think that the self-driving feature is going to save more lives than seat belts airbags and every other thing that we've done to cars combined because that's that's the way to go absolutely i mean worldwide there's about a million people a year dying in car accidents and that's a hell of a lot and

48:21yeah so there's like 10 million people that get seriously injured so it's like you know we got to hustle on this yeah and and why not we got the technology to do it yep anyway thanks very much elon all right anyways i'm i'm so flattered i'm so happy i got a chance to meet you all right talk to you shake your hand thank you so much [Music] you

资源与出处

相关视频

同频道、同主题或相邻年份