大会演讲·中文精读

百度CEO李彦宏对话比尔·盖茨与马斯克(2015博鳌论坛)

Baidu CEO Robin Li interviews Bill Gates and Elon Musk at the Boao Forum, March 29 2015

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

2015 年 3 月 29 日,博鳌论坛。李彦宏坐在中间,左手边是比尔·盖茨,右手边是埃隆·马斯克。三个人凑在一起聊了一个小时。

李彦宏开场说,两周前答应做这场对话的主持人之后,他在网上征集问题,收到了 500 多个。他挑了两个,一个给盖茨,一个给马斯克。后面的问题则从技术、慈善、政府治理一路滑到自动驾驶和 Uber。

这场对话信息密度不低。盖茨讲了他对"野心"的真实看法,马斯克承认 SpaceX 和特斯拉最初的存活概率只有一成,两人在人工智能安全性上站到了同一边,又一起给中国政府主导的 AI 项目提了反对日本式做法的建议。十年后回头看,很多话都说中了,也有一些明显没算准。

下面按话题拆开讲。

"想超过我成为世界首富":盖茨说别把目标定成这个

第一个问题来自一位自称"有野心的年轻人":我想超过你成为世界首富,需要什么样的愿景,给点什么具体建议。

盖茨的回答绕了一下。他说,衡量卓越没有单一指标。接着他讲了自己的运气:出生在一个计算正从集中、昂贵走向分散的年代。那时候软件在产业里根本不算核心,硬件才是主角,大家忙着买机器、维护机器、伺候机器。微处理器出现之后,计算变便宜了,存储也变便宜了,他才有可能把软件想成"给个人的工具"。

他举了一个具体例子。当年软件行业的销售冠军奖,标准是卖出一万份,能拿一个大奖杯。微软的 BASIC 卖出去了 600 万份。他打电话过去问,六百万份你们有什么奖。对面愣了:你说的是什么软件。

我从来没有坐下来对自己说,好,我要比 IBM 更强。我没有那位年轻人那样的自信。我只知道我喜欢软件,想折腾它,然后它长成了一家公司。(比尔·盖茨)

这一段值得细看。盖茨说,他管理公司时的心态是:也许这家公司会变大一倍,那我先为"变大一倍"做好准备。他刻意没对自己说过"这家公司一定会变得巨大",因为一旦这么想,人容易跳过那些真正困难的扩张工作。等他管着十个人的时候,如果有人告诉他"你将来会有十万员工",他会笑出来,然后反问:十万人能制造多少 bug,我不想要十万人。

李彦宏追问了一句:所以你的意思是,不需要那么有野心也能成功?

盖茨的回答是:我不认为我们中的任何一个人当年就笃定这些事会成。如果你一开始就知道没风险,我不确定你最后真能走到那儿。

盖茨眼里成功者的共同点:近乎偏执

被问到"你还会看重哪些特质"时,盖茨讲了一段往事。

他参加过一个行业论坛,台上七个人辩论:未来的计算机界面是字符模式,还是图形界面。当时图形界面慢得可笑,给它写软件体验极差,那是 Windows 1.0 的时代。台上大部分人认为搞图形界面是蠢事。盖茨坚持说这东西会成。辩论到最后,台上另一个人说了一句话:比尔错了,但比尔比我们所有人都努力,所以哪怕方向错了,他也更可能成功。

盖茨说,他收到过的最好的赞美就是这一句。

那个阶段我很偏执。我不相信假期,不相信周末。结果这套对我有效,我们让公司跑得比其他人快,犯错误也比别人快,然后更快地看到错误。(比尔·盖茨)

他还补了一条公司层面的选择。当年大部分软件公司把自己定义成"单一产品公司",从组织结构到销售预期都按单一产品设计,而且默认主要市场在美国。微软从一开始就规划做多产品公司,第一件产品 BASIC 解释器跑在一堆不同机器上。他们的自我定位是"世界上最有效率的软件工厂",招软件工程师,写大量的软件,而不是某一件软件。

马斯克后来接了一句玩笑:盖茨的软件工厂不用买设备、不用盖楼,马斯克得真盖厂房、真买机器,所以特斯拉搞了个 Gigafactory(超级工厂)。盖茨回应说,他们搭的是一个由编译器、符号调试器这类东西组成的"软件超级工厂"。

马斯克:我一开始就认为大概率会失败

第二个问题同样来自网友:马斯克怎么能有那么多精力做成这么多事,他是按步骤推进,还是一开始就设定了宏大目标。

马斯克的回答很直接:回到比尔说的那点,你其实并不预期事情会做大。

一开始我预期的是失败,那是最可能的结果。SpaceX 大概只有 10% 的成功概率,特斯拉差不多。你看看航天创业公司和汽车创业公司的历史记录,相当糟糕。(埃隆·马斯克)

他说,如果按"投资回报和风险的比值"给所有行业排序,航天和汽车基本在最底端,那是片坟场。2008 年底,特斯拉、SpaceX、SolarCity 三家几乎同时濒临死亡。

他做这些事的动因,来自他觉得有几个重要问题需要被解决,而看起来没人在解决。"我要做一家大公司"这个念头并不在起点上。特斯拉和 SolarCity(家用太阳能公司)指向同一件事:加速可持续能源的推进。至于电动车,他原话是,当年电动车被普遍当成最愚蠢的东西,"愚蠢的汽车、愚蠢的太阳能",他们好几次差点没挺过去。

李彦宏追问:那你怎么能同时管这么多事?

马斯克的回答里有一句很少见的坦白。

我不推荐这么做。这对生活质量真的不好。我管 SpaceX 和特斯拉,但我不管 SolarCity,我只提供战略指导,那边有人做得非常出色,我大概一个月花一天。特斯拉和 SpaceX 是每天、每个周末、每个假期。(埃隆·马斯克)

他专门强调了团队的功劳:大部分时候外界给了他太多赞誉和关注,因为他是那个"看得见的部分",公司能成,全靠里面有极强的人在。

盖茨:100 年后人们记住我什么,我不在乎

李彦宏把话题转向盖茨:网上关于你的问题大多和微软有关,你现在花在基金会上的时间比微软多,你觉得通过基金会对人类的贡献,比通过微软更大吗?换句话说,一百年后,你更希望被记住是微软的创始人,还是伟大的慈善家。

盖茨先开了个玩笑:我希望一百年后我还活着,那就不用在乎别人怎么记我了,只要还能做有意思的工作就行。

然后他给了自己的判断框架。经济里最大的一块是私营部门,二三十年后世界变得更好,绝大部分功劳要归给私营企业。数字革命还在早期,像百度、谷歌、苹果、微软这样的公司,它们造出来的工具被用在无数领域。他举了教育:我们还没看到,但早晚会看到,这些工具能把一个人正在学的东西和他的目标连接起来。

任何有机会参与这件事的人,尤其是建立起一家像样公司的人,都做出了巨大贡献。而我看到的是,我在五十多岁的时候最独特能做的事,是全职投入基金会。(比尔·盖茨)

他说,微软有的是优秀的人可以接手。他自己二三十岁的时候有个观点:CEO 不该超过 45 岁,因为谁能跟得上这些变化。等他到了五十多岁还坐在那个位置上,他对自己说,我不该还在这儿。

基金会那边,他和妻子梅琳达一起做。他当时的想法是:谁去研究疟疾?这类问题藏在贫困国家深处,大多数人不会亲自去看营养不良和传染病造成的损害。组织科学家、组织政府去攻这些难题,需要的正是他在软件革命里练出来的那套本事:带团队、挑科学家、对需要十到十五年才有突破的事情保持耐心。

"能发现自己最大的贡献和最大的乐趣是同一件事,这挺好的。但我二十多岁的时候,做这个不会满足。"

他补了一句关于形象的话:人们记住的更多是微软,基金会没那么显眼,大部分工作在穷国。他正在推动和中国的关系,因为中国有大量创新可以帮助非洲,无论是健康、农业还是手机相关的技术。

人工智能:马斯克说这是"召唤恶魔",盖茨站他这边

这场对话真正有火药味的部分在这里。李彦宏提到,前一晚他和盖茨就人工智能有过一场小辩论。他还提到马斯克最近说过,人工智能的进展就像"召唤恶魔",这话引发了激烈争论。而某位首席科学家在采访里说,担心人工智能的阴暗面,就像担心火星上人口过剩。

马斯克的回应不留情面:这个类比完全不准确。

数字超级智能的风险在于,它不会停留在人类水平,会达到超级人类水平,远超任何人类,远超你能想象的任何东西。(埃隆·马斯克)

他用了一个类比来说明问题的性质。核研究有非常危险的武器潜力,麻烦的地方却不在释放能量,释放能量很容易,难的是安全地约束住这股能量。所以 AI 研究的重心应该放在 AI 安全上,投在安全上的资源应该远远多于投在推进 AI 本身上的资源。

"它可能是好的,也可能是坏的,而且可能是灾难性的坏。如果真的出现某种意义上的堆芯熔毁,你一定会希望事先做了安全设计。我不反对发展 AI,但我认为应该极度小心。如果这意味着 AI 要多花一些时间才能做出来,那这个交换是值得的。我们不该一头冲进自己不理解的东西里。"

李彦宏转向盖茨:我知道你持类似观点,你和他之间有分歧吗?

盖茨说没有。他提到马斯克已经投了钱去推动这方面的研究,这非常好。他推荐了一本书,Nick Bostrom 的《超级智能》(Super Intelligence),然后讲了他自己的理解版本。

进化给了我们一个通用的学习算法,但它跑在一台极慢的计算机上,内存极小,向其他计算机传输数据的带宽也很有限。我们脑袋上这个玩意儿,每次新建一个,它连走路都不会,要花很长时间学。所以相信我,一旦这个"把经验转化为知识"的算法被做进软件,一旦它被植入硅片,你不一定能察觉到它什么时候跨过人类水平,它几乎会立刻到超级智能水平。(比尔·盖茨)

他补了一层:随着时间推移,那块硅片会越来越强,而一旦它有了学习算法,它就会跑到互联网上去读所有的杂志和书。我们其实一直在给超级智能搭建内容基础。你以为你在用互联网,实际上你在给未来的超级智能准备语料。

盖茨说他自己尽量不去激动,但当有人说"这不成问题"的时候,他就开始真正担心了。他原话是:怎么可能有人看不到这是多大的挑战。

李彦宏的"中国大脑":盖茨拿日本当反面教材

李彦宏接过了话头。他说自己最近在读 Walter Isaacson 的《创新者》(The Innovators),书里详细讲了创新是怎么发生的。他的一个体会是,创新通常来自许多有才华的人的集体努力,但另一个体会是,美国政府尤其是军方在技术推进中扮演了重要角色。阿波罗计划给微芯片创造了巨大而稳定的需求,没有它就没有英特尔这类公司的繁荣;互联网则直接脱胎于 ARPANET。

他接着说,这个月早些时候他在政协会议上做了公开演讲,倡议一个叫"中国大脑"的项目:由政府出面建设人工智能基础设施,建一个足够大的平台,让公司、创业者、大学都能在上面做研究。他问两位有什么建议。

盖茨先肯定了方向:中国加大基础研究投入是好事,无论医学还是计算机科学,这会强化大学,会加快创新。然后他立刻给了警告:政府与创新的关系怎么设计,要非常小心。

他举了日本的例子。日本当年选了非常具体的目标和路线,就是所谓的第五代计算机项目,那条路走错了。八十年代美国一度觉得日本比自己强:制造比我们强,规划政府 AI 项目比我们强,高清电视比我们强。但盖茨说,那恰恰是美国非常好的一段时期,因为美国人被打击了,谦虚了,然后回头去看自己那套"分散、没有中央计划"的体系。八十年代最好的工作,包括 PC 和互联网的基础,都是由一个高度分散的系统完成的:资助几十上百种不同路线,办竞赛,比如 DARPA 的无人车挑战赛,现在还有 DARPA 机器人挑战赛。

我强烈建议不要做日本做过的事,去看一看为什么美国这套模式,十年又十年地跑下来,一直是最好的。(比尔·盖茨)

他还专门绕开了一个例外:曼哈顿计划是特殊情况,那次确实按那种方式做成了。但美国做计算机科学的方式跟曼哈顿计划完全不同,是同时押注几十种不同方案。

马斯克同意,并补了另一个角度:重要的是创造一个培育创新的环境,让它以自然的方式演化,而不是在高层选定一项技术,认定它就是赢家,因为它可能不是。他强调了一条具体的机制:对失败的惩罚要低。

硅谷能成,一条重要原因是,很多创业公司非常成功,也有很多失败了,但人们很快重组,离开,加入别的公司,这套回收机制非常好,是一种递归的自我改进。你要支持创新和新技术,就得承认路径是未知的,没有地图,没有路标,你一定会走错几步,所以犯错必须是没关系的。(埃隆·马斯克)

两个人各自学到东西的人

李彦宏问盖茨:你成名比我们早很多,应该有机会认识上一代企业家,有没有谁是你尊敬、并且从他身上学到东西的。

盖茨列了一串。IBM 是他创业时的主导公司,那种"庞然大物",1981 年微软和 IBM 一起做出 IBM PC,跟 IBM 合作本身就很奇妙,他从 Frank Cary 这些人身上学到很多。英特尔那边,他从 Andy Grove 身上学到很多,他补了一句"Andy 对我不总是友善,但那不妨碍我学"。他推荐 Grove 的管理书。Gordon Moore 在芯片层面的技术贡献惊人,现在也是了不起的慈善家。Steve Jobs 是个不可思议的人,盖茨的评价角度很有意思:

乔布斯之所以如此不可思议,一部分原因是他极其擅长打造伟大的产品,尽管他对工程了解不多。他挑人的能力,以及他在并不真正理解软件或芯片规则的情况下定义产品的能力,让他更了不起。(比尔·盖茨)

然后他说了一句出乎意料的话:他可能从 Warren Buffett 那里学到的更多。Buffett 完全在这个疯狂的计算机行业之外,但他对商业的基本看法、对商业模式怎么运转的看法,对盖茨的影响比任何单个人都大。

还有一条来自客户。八十年代日本是微软的重要客户,头几年日本占了微软销售额的 40%。盖茨说,日本客户对质量的强调把微软叫醒了,改变了很多做事方式,是一次巨大的进化推力。

李彦宏追问:除了技术圈之外,还有别人吗。盖茨说,现在进入慈善阶段,他开始认识很多不同行业的创业者,但当年有人说"你是企业家",他总是反问:那是什么意思,我是个搞软件的。他十二岁写软件,十三岁写软件,从十一岁到十七岁一直痴迷这一件事,那是他唯一有可能做到世界级的事。他并不可能去做别的行业的公司。

"然后我得学着雇人,天哪,还得学着裁人,那可太有意思了。得做预算,得卖东西,得去展会站在展台前,脚都站肿了。我只是想给自己做一个能让自己更自主的软件,然后我得学会合同,学会一大堆配套的东西。所以某种程度说,那股劲儿全来自对软件的梦想,跟'我要开一家公司,随便挑个行业干一票'没什么关系。"

代际差异:盖茨说这一代更"温和"

被问到不同代创业者之间的差别,盖茨说他很欣赏马克·扎克伯格,也和他熟。他注意到这一代有一个优点:很善于向老家伙们提问,"你犯过什么错误""这事应该怎么办"。

他描述自己那一代的环境:他们是在一张干净的白纸上作画。当年去任何会议,满场都是同龄人,没有更年长的人,分析师和风投也都很年轻,那是一个二十多岁的人组成的行业,后来变成三十多岁。现在年龄结构多元多了。扎克伯格当时三十出头,已经是有超级成功公司的年轻人里最年轻的一批。

我觉得这一代可能更温和一点。乔布斯在某种意义上是疯狂的,我在某种意义上也是疯狂的,也许现在更难那样了。(比尔·盖茨)

马斯克说他不好概括,因为性格差异太大。他提到自己认识的几位中国企业家,包括 Jack Ma(马云),但他加了句实话:人们互相见面的时候通常都表现得最好。他真正想看的场景是工程会议:有人进来说这东西不工作了、要延期了,或者说这实在太慢了、我们是不是得从头重建。那种时刻他想在场,看这些人怎么处理挑战。因为所有生意,从外部看和从内部看是两回事,早晚会走到必须重新审视当初假设的那一点。

盖茨接了一段很有分量的回忆:微软每次更成功,就有人说"这是他们最后的成功了"。这话让他很不舒服,因为当别人说"他们迟早会出错"的时候,他知道那是真的。他小时候崇拜的那些公司,Wang(王安电脑)和 Digital Equipment,早就没了。看着它们失败对他可能是好事,那说明:卓越的技术和优秀的人才,完全不能保证你不被市场转向甩下。

大学时代的叛逆,和"79 个冒充者"

李彦宏提到盖茨大学时期的往事:只旁听不注册,连续工作三十六小时,睡十小时,醒来接着干。他问这算不算成功者必备的东西。

盖茨的回答很有意思。他说他和乔布斯各自以不同的方式"有点疯",但"我没像他那样吃那么多致幻剂",而且他不认为这是必要条件。有人争论乔布斯如果对人好一点还能不能伟大,盖茨的答案是能,"他确实伟大,我不想抹掉这一点,但以我的情况看,有些阶段我对人非常苛刻,大概因为我对自己非常苛刻,然后你得学会别把这种苛刻投射到管理风格里,这其实是很自然的反应"。

他还讲了哈佛数学课的故事。他在中学里是那个"数学好的人",到了哈佛,班上有八十个人,每个人的自我定位都是"我这辈子见过的最擅长数学的人",于是其中七十九个是冒充者。他最后考了第二名,第一名后来成了纽约很成功的律师。

至于不去上课,他的解释很坦率:那只是要有一个独特的定位,让自己和那八十个人不一样。

马斯克给创业者的建议:别想着创业,想着做出有用的东西

马斯克在这里讲了一段几乎是整场对话里最实在的话。

与其说"我想当创业者"或者"我想赚钱",不如问:有什么有用的东西是你希望存在、但现在还不存在的?然后试着找其他人一起把它做出来,并且不断做得更好。如果你真的创造出了有用的东西,钱自然就来了。这是运转正常的经济体奖励有用商品和服务的方式。如果你做的是自己热爱、并且你认为别人也会热爱的东西,你就更容易牺牲时间、投入精力,而且即使没成,你也不会后悔。(埃隆·马斯克)

他顺带讲了自己怎么学会造东西。他说自己原本不会制造,没见过 CNC 机床,不知道怎么铺碳纤维,这些都不懂。但只要你读书、和专家聊,很快就能上手。他认为人们严重低估了自己的学习能力,这件事其实相当直接:读书就行。

他补了一个关于阅读速度的观察:一个人说话的输出速率大概是每秒几百比特,你听的时候能处理的信息量只有每秒几十比特,而阅读能处理的是这个的好几倍。所以听有声书或者听人念教材的时候,他会觉得"这太慢了",像在听人把课本念给你。

工程师治国 vs 律师治国

李彦宏把话题拉到治理。他提到刚去世的新加坡建国总理李光耀,提到中国领导人中有不少工程技术背景,倾向于用工程化的方法治理国家,这和美国很不一样,美国"基本上是律师在治国"。他问两人怎么看这两种体系的利弊。

马斯克先答,很直接:作为工程师,当然偏向工程化的方式。美国律师太多了,有些人不该是律师。

总体上我认为技术专家主导是好事。我希望美国国会有更多懂科学的人。(埃隆·马斯克)

盖茨的回答更长。他说自己有幸和李光耀吃过几次饭,印象是李光耀不只是深思熟虑,他敢于从西方体系里挑一部分,说这部分好、这部分未必到处适用、这部分我不同意,这在当时是很胆大的,因为西方体系正在成功,世界上所有富国走的都是西方那套。李光耀敢说"我要做得稍微不一样",本身就是巨大贡献。盖茨也承认新加坡是城市国家,它在政府高薪这类事情上的做法不一定能放大到更大规模。

然后他给出了自己的框架:真正想要的是民主和专业知识的混合,而没有任何国家把这个平衡调对了。偏向民主那一侧,会有极端情况,比如选上来一个错误的人,以及怎么把他换下去、怎么让新人上来,这些机制有巨大优势,帮了美国很多。但到了复杂问题上就露怯了,比如怎么高效运营医疗体系,为什么美国花了那么多钱,你找不到一个民选代表能跟你认真讨论这个系统的动力学,讨论它为什么和其他国家不同、该怎么改。

他对比说,中国政府的关键岗位上有大量工程师和科学家,愿意看别的国家怎么做,而且有一个做法他很欣赏:新政策先在一部分地区试点,看行不行,再推开。他说中国政府更像是在"研究政策",而英国议会或美国国会那种互相喊"我对""你不对"的场面,在这里不常见。

我从没坐在美国国会里听人说,来,你这个方案在一个州试试,我这个方案在另一个州试试,然后我们把最好的组合起来。我们现在典型的选举对话不长这样。(比尔·盖茨)

他划了一条界线:有些事情比如大学怎么办,美国模式别人直接照搬就行;但医疗体系和治理这类事,也许应该取一部分,然后试不同的变体,让 192 个国家(包括次国家层级)的不同经验产生价值。

强政府会不会压住创新

李彦宏追问:中国政府执行力很强,基础设施比如高铁、高速公路建得飞快,可能是世界上最大的交通基础设施。但强政府会不会在某些地方管得太严,压住创新。

盖茨的回答分了领域。他认为中国的创新速度已经很快,政府还可以把科研投入再提高,而且正在提高。有些领域确实受影响,比如媒体,中国市场在这一块一直落后,这部分国家大概要负一点责任,不过情况在变。但在能源这类领域,国家政策没有挡住谁去搞大发明。

他举了自己投资的公司:TerraPower,做下一代核能。他说中国是他们最自然的合作伙伴,因为中国很像六十年代的美国,那种"要往前走、要做新事情"的意愿非常明确。

今天美国非常谨慎,我们对现状挺满意,所以如果有人想盖新东西或者采用新方法,会经过五道关卡,每一道都可能是"不行"。而中国有一种"往前做新事"的偏向,这有小的下行风险,但上行空间巨大。所以像核能这样的突破,更可能出自中国,而不是几乎任何其他地方。(比尔·盖茨)

他补了一段技术史:五六十年代是美国在干这个大工程,七十年代之后是日本,后来韩国接过了这个角色。那种"偏向做大工程"的劲头对世界是好事。

五十年后还要不要开博鳌

李彦宏问了个轻松点的问题:五十年后,像博鳌这样的会还要办吗,人们还得飞到同一个地方面对面,还是会有更好的沟通方式。

马斯克说,机场会更近一点。他认为人们还是会想开会,哪怕只是为了情感价值,靠近别人就是感觉更好。就算有完美的虚拟现实,戴上头盔、穿上触觉服,感觉"真的在那儿",那和真的在那儿仍然不是一回事。

自动驾驶:技术五年,法规再加两三年,整车替换要二十年

全场最具体的预测在这一段。

马斯克先给了一个量的框架:全球路上大概有 20 亿辆车,年产能大约 1 亿辆。所以哪怕明天就出现完美的自动驾驶汽车,替换掉这个庞大的工业基础也需要很长时间,大概二十年。对经济学家来说这个数字可能小一点,但仍然要很久。

然后他分层给时间:

  • 技术层面,能做出比人类司机安全得多的车,他认为还需要五年以上。
  • 技术做出来之后,还需要监管者认可它真的安全。他预计会有两到三年的延迟。
  • 他提到一种验证方式:先让自动驾驶系统在"影子模式"下运行,人在开车,系统在后台记录它本来会怎么做,和人实际操作对比,用统计证明它更安全。
  • 整车保有量的替换,二十年量级。

盖茨补了一个关键视角:这件事不会一刀切。他建议世界上有几十个司法辖区先跳出来,"你们先上,先拿好处,也先承担问题",这样就能积累几亿小时的软件运行数据,让软件见识"球滚到马路上"这类场景。先上的地方会有代价,但这正是没有单一全球政府的好处。他甚至说,哪怕在中国,也应该挑几个地方,给超级宽松的规则试。

我们应该有几十个辖区说,来吧,就这么设计,十年后我们既看到好处,也看到几个问题。(比尔·盖茨)

他还提了城市设计层面:如果要设计一座假设无人机可以送货、假设车能自己避撞、假设不需要停车楼的城市,那就该从头设计。我们没有这种奢侈,绝大多数情况是在现有巨型城市里改造,问题是怎么平滑过渡。

盖茨在这里把 Uber 拉进来当例子。他说 Uber 某种意义上就是一种"无人驾驶车",只不过用的是按需的人类劳动力。真正的变化在于:过去"什么时候该用一件资本品"这件事,市场机制处理得非常低效,现在有了数字平台,需求方和资产方的匹配摩擦大大降低,这会改变一大堆关于资本品和劳动力使用方式的东西。

李彦宏问马斯克:你把 Uber 看作伙伴还是竞争对手?没人需要开车之后会怎样?

马斯克的回答很干脆:伙伴或者对手都行,我们只是造车,我们可以把车卖给 Uber,别人也可以买。车会逐渐变得更自动。但他再次强调工业基础那条:这件事不可能一夜之间完成,因为保有量是 20 亿辆,产能是 1 亿辆。

李彦宏最后问:那司机是不是该马上开始恐慌。马斯克的回答是:谁知道那种环境下会发生什么。

这场对话里最值得拿走的几条判断

十年后回看这场对话,有几个地方值得单独拎出来。

盖茨和马斯克对"野心"的看法高度一致,而且都反常识:他们当年都不认为事情会成,盖茨甚至刻意不让自己相信公司会变得巨大,因为那种信念会让人跳过必要的苦活。流行的"先定一个宏大目标"叙事,在这里被两个人同时绕开了。

两人对 AI 的判断在 2015 年就已经很清楚,用的还是同一个逻辑:争论的重点在"超级智能和人类的水平差距太大,安全投入必须走在推进之前","AI 会不会变坏"这个话题被放在了后面。盖茨那句"我们一直在给超级智能准备语料",放在大模型时代看,几乎是字面意义上的预言。

关于政府主导创新,盖茨给的日本第五代计算机案例,是这场对话里最有操作价值的一段。"政府该不该投"这个问题他并不纠结,他把注意力放在了另一处:政府该不该指定路线。他的方案是投钱,但分散押注几十条路线,办竞赛,让失败便宜。

至于自动驾驶,马斯克分层给的时间表,技术五年、法规两三年、整车替换二十年,十年后看,技术那一层比他自己估的乐观了些,法规那一层大致在节奏上。他反复强调的"工业基础决定速度",是很多人容易忽略的约束条件。

最后给几条可以直接拿走的行动建议:

  • 做产品时先问"有什么有用的东西是我希望存在但现在没有的",把"我要创业""我要赚钱"这两句从脑子里删掉。钱会跟着有用性来,起点别放在钱上。
  • 如果准备押注一个新技术方向,先确认自己在做"分散押注"还是"选定唯一路线"。前者靠组合,后者靠运气,政府和大公司都容易犯后者的错。
  • 设计任何鼓励创新的机制时,把"失败的代价"当成一等公民来设计。硅谷真正强的地方在于失败了能快速重组,失败本身从来不是它要消灭的东西。
  • 评估一项技术落地时间,至少分成三层:技术可行、监管认可、存量替换。大多数预测错在只算了第一层。
  • 想在某个新领域做事,别用"我没学过"当理由。马斯克的说法是读书加找专家聊,这件事比大部分人想的直接得多。背景没那么稀缺,判断力和耐心才稀缺。
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)。

0:00[Music] so good morning everyone and I'm prly I'm the chairman real Pleasant to be the monitor of this extraordinary panel in the middle we have Bill Gates the co-chairman of the m g Foundation former chairman and CEO of Microsoft is still the technology advice microp right and on the left we have Elon mask

1:00he is CEO and CEO of SpaceX and CEO and Chief technology architect of Tesla and chairman of so welcome both so roughly two weeks ago when I agreed to be the moderator of this this panel I I posted a a message by Dua sisting questions from the internet users I got more than 500 questions for two of you and I'm going to select one question for each and on behalf of the users I'd like to to ask

1:54so question so for Bill we have a question from a very ambitious person he said as a young man who intends to overtake you as the world's richest man what kind of vision do I require and what specific advice would you give me well fortunately I don't think there's any single metric that defines Excellence you know hopefully that person has some ideas where he sees beyond the current generation my luck was to be born at a time when Computing was going from a a very centralized expensive form where the amount of software that was needed was very very modest and so it wasn't even considered the magic element of the system the hardware just keeping it

2:50running and buying it and babysitting it was the center of the industry and yet because the micr processor coming along and in essentially from my point of view making Computing free and storage free and eventually even at gam I could conceive of a form of software in terms of a tool for the individual that was radically different than any other software I I remember at the time the industry award for the top selling software was that you sold 10,000 copies got their big trophy and they gave you a big b to it well our basic had sold 6 Mill copy so I I called them up and I said what do you have for 6 million they were like what what kind of software are you

3:37talking about and so it was really utterly different you know for this gentleman hopefully at a young age he'll see how to make robots or how to make Assistance or some way of making software development far more efficient than it is today you even though so many things have changed if you look at you know say the code that defines the website or the code you write for business applications I have to say I would you'd asked me four years ago what that would look like I would predicted that it would be much easier to do business applications much easier to write highly reliable code so there there's an opportunity for somebody and it's not just in software if they can invent you

4:25know a way of making energy cheap and CO2 free see a Twist of using technology to do that I'd say that would be a kind of a mindblowing contribution so the room for Innovation the speed for Innovation is better today than ever in my case I didn't sit down and say okay I'm better than IBM or something I didn't have the same confidence that this person had I knew I love software I wanted to mess around with it and it led to you know

4:59being able to create a company but I didn't know at at any point in time part of the reason I think I ran the company well so I thought okay maybe the company will be twice as big let me get ready for that I never said of course will be gigantic because then you might skip all the hard work of scaling up that factor of two and I thought this is it you know I just have to build a management system for 5,000 people if you said to me when I had 10 people I on with code hey you'll have 100,000 people I would have laughed I said wow how many bugs can you create with 100,000 people I don't I don't want 100,000 people so you know I I wish good luck so you're saying you

5:41don't really need to be this ambitious in order to be successful right well I don't think I we can ask Elon I don't think any of us assumed that the things would line up to create you know in his case multiple successes that are are pretty fantastic I think if you had that you knew that you didn't think it was at risk I'm not sure you would you would actually get there what other characteristics would you look for for potential successful well they're you know people who end up being successful and are are often not always pretty fanatical about the things they're trying to do I remember one industry panel where there were about seven people and the debate was would the

6:37computer interface be this character mode thing or would it be Graphics user interface and at the time the graphics user interfaced up was so slow it was laughable and writing soft for it was so bad it was Windows 1.0 and so the people on the panel were saying no no this this is kind of a stupid thing that I was say no no believe me this will be great and one of the guys on the panel said that that hey Bill is wrong

7:07but Bill Works harder than the rest of us so even though it's the wrong solution he's likely to succeed and that was the best compliment I ever had was just by working day and night I could send the industry in some funny Direction so you know I was fanatical in that period of time that is I didn't believe in vacation I didn't believe in

7:29weekends you know so it turned out that work for me that we got our company going at a speed that allowed it to make mistakes faster than than other people were and kind of see those mistakes and many software companies at the time thought of themselves as a single product company and they were literally designed to be a single product company and they thought that their sales were mostly going to be in the United States we from the beginning planned to be a multiproduct company you know based I don't people have ever heard of that was our basic interpreter that ran on a bunch of computers was our very first product and of course today they're

8:09still around but you know people use JavaScript there a lot of the things that the basic was used for them so we had a conception of a very big company and we knew what we were about with software hiring software people and writing lots and lots of software not a specific piece of software so that helped us a lot we said we're going to have the world's most efficient software Factory Elon has to actually build buildings and buy equipment to make the gig we set up gigafactory of software which is with a bunch of guys with better symbolic some better cers and compilers and things like that and we you we didn't get other companies were were also doing that but we were

8:54somewhat ahead in terms of being serious about high quality software development all right question for ELA also from our internet users I'm so curious that I'm going to explode how does must have the energy and inm to accomplish so many great things it's incredible it's just too cool does he do things step by step or does he have huge goals already set for projects that have just hatched sure well I think first of all it's

9:30back to what B said you don't actually expect things to be big in fact beginning I I expected that I thought we fail that was the most likely outcome by far actually maybe 10% chance of success was space something similar for Tesla if you look at the the track record of you know space startups and car startups it's pretty bad so some from a standpoint like wow you know if you're to rank order and say what what I'm company what's like the highest ret investment for the risk

10:06like the right at the bottom of the list of of the least desirable would be something in space and and cars you know it's it's not it's it's a graveyard so the don't want be repl at this point but came very close to to both Tes basx and Solar City all all dying at the end of 2008 so yeah it's but in terms of how it's it's really just that I thought there were these important problems that needed to get solved and it do seem as though anyone was really be solving them you know with respect to Tesla and solver City it's about trying to accelerate the advancement of sustainable energy sustainable energy production Sol City sustainable energy assumption in the case of of

10:55Tesla and we when we started Tesla electric cars were just sort of as the stupidest thing really thought stupid car stupid square and narly escaped that so many times so it's not there wasn't some grand plan the expectation was that that was the most but the the followup question would be how can you manage so many things to nend fre compan well I don't I don't recommend this I really it's not good for quality of

11:43life I should I should EMP I run SpaceX in Tesla but I don't run Silver City I provide strategic guidance fortunately there does an amazing job it's like one day a month that thing T and SpaceX are kind of every day every weekend every every holiday not just as the CEO you also in charge of Technology product you a very ended person we have

12:12like we have really talented team there so I just want to emphasize that sometimes I get in fact most of the time I get way too much credit or attenion for what I do I just like the visible element but the reason those companies are successful is because we have extremely people at I really don't think it's a good idea all and yeah I mean I do it because I I feel like I'm responsible for these things not because it's necessar the best thing quality okay yeah got it so among the

12:50five kind of questions when I went through those questions one thing I noticed is that for questions to build most of the questions are Microsoft related and I understand you spend more time on your foundation than Microsoft this days do you really think you are making a bigger contribution to humanity your

13:18foundation through your foundation than you did through Microsoft in other words let's say 100 years from now would you be rather be remembered as the founder of Microsoft or as a great philanthropist well I hope I'm alive 100 years from now that be they don't have to remember me as long as I get to do fun work like I am now you know the the biggest part of the economy is the private sector and when you think you know why are things going to be better 20 years from now 30 years from now the vast majority of that's going to come private sector companies what the digital Revolution is at an early stage and so companies

14:07like Buu Google Apple Microsoft they're making a huge contribution the tools that they create that get used in so many domains you know just take education as one domain that over time we haven't seen it yet but over time that should make ability to connect up to what somebody's learning and what their goals and aspirations are so this is incredible stuff and anybody who had a chance to be part of that particularly to build a good siiz company made a a huge contribution what I saw was

14:46that my the most unique thing I could do starting in my 50s was to be full time at the foundation that is that there were other great people who could take the mantle of Microsoft when I when I was running Microsoft my 20s and 30s I I always thought you know for sure it should never have a CEO over 45 because who could keep up with all this stuff so then you know as I found myself my 50 still company was like I told myself I shouldn't still be here and anyway the foundation work which is in partnership with my wife Linda became more appealing and I was saying myself gosh who's going to work on malaria it it appeared to be more of a back that organizing science and

15:37government to take these proms that are a little bit hidden in way in these core countries where most people don't actually go and see the the depredations of malnutrition and infectious disease and so it feared like an opportunity that showed up that could take all the skills of organizing people and energizing people in the software evolution I'm now in a slightly different domain but but the skills of running teams and picking scientists and being patient about things that take 10 or 15 years to have a breakthrough I could take all of that and so I'm I think for this days of my life this is my biggest contribution and the most fun it's nice when you decide those two

16:20things are the same but in my 20s I would not have been satisfied at doing this and so I think in sh the image is more about the Microsoft the foundation isn't that visible most work of the foundation is in poor countries we are trying to build more and more relationships because there's so much innovation in China that can help in Africa whether it's in health or in culture or cell phone magic so

16:51that that's a growing thing so you should come to China more often I do so you you and I have had a short debate on artificial intelligence last night and I want to bring that topic up here too I also understand recently you said that artificial intelligence advances is like summoning the demon that generated a lot of hot debate by those Chief scientist Ang recently said during an interview that worrying about the Dark Side of artificial intelligence is by woring about over population non Mars how would you address that he he said it it's it's a distraction to with the scientist being artificial intelligence I I don't think I think that's radically inaccurate analogy the

17:48I know I know bit about so the the risks with digital super intelligence it's important to appreciate that it would be it wouldn't be just human level would be super human almost just zast humans to be Way Beyond anything really imagine I think more be like if you instead of saying nuclear research with there potential for a very very dangerous weapon you releasing the energy is easy that containing that energy safely is very difficult and so the the right I think the right emphasis for AI research is on AI safety we should put vastly

18:40more effort into AI safety than we should put into advancing a in first place U because it it may be good or it may be bad and it could be catastrophically bad if there could be the equivalent of meltdown so you really want to EMP safy so I'm not I'm not against the Advan this I want to really about this but I I do think we should be extremely careful and if that means it takes it longer to develop AI then I think that's that's the right tradeoff we shouldn't be rushing head into something that we don't understand B I know you share similar views with but is there any difference between you and him I don't think so I mean he actually put some money out

19:30to help get some work going on this and I think that is absolutely fantastic you know for people in the audience who want to read about this I highly recommend this Bost from book called Super intelligence and but the basic point that Elon just made that we have a general purpose learning algorithm that Evolution has endowed us with and it's running in an extremely slow computer

19:57very limited m size ability to send data to other computers we have this m thing here and whenever we build a new one starts it doesn't know how to walk really long gr process yeah so believe me as soon as this algorithm of taking experience and turning it into knowledge which is so amazing and of course we have not done in in software as soon as you do that it's not clear you'll even even know when you're just at the human level you'll be at the superum level almost assum as that algorithm is implanted in Silicon and you know actually as time goes by that silicon piece ready to be implanted the amount of knowledge as soon as it has that learning Alum it

20:50just goes out on the internet and reads all the magazines and books and things like that we have essentially been been building the content based for the super intelligence you think you're using the internet that's actually what you're you're doing so you know I try not to get to exercise about this but when people say it's not a problem that really then I can start to really

21:16in a point of disagreement how can they not see what a a a huge challenge this is yeah we we are understand this will have huge impacts in the society in the future and many companies include Microsoft by we all investa in artificial intelligence during a this days I'm reading a book by walk is the

21:47innovators there are lots of detailed stories about how Innovations came out I guess his point is that Innovation come from usually from Collective effort from many talented people but what I also took away is that the US government especially the military played an very important role in the advancement of technology for example project Apollo that generated a very large and steady demand for microchip which without it in the the flourishing of catic companies like Intel and internet was a a

22:37direct output of of ARA net so the US government played a very important role in this kind of technology advancements earlier this month I was at the CPCC meeting in China I I made up a public speeech advocating a project China brain it's basically a artificial intelligence infrastructure put out by the Chinese government we're hoping to build largest AI infrastructure so that companies entrepreneurs

23:16universities can do research based on this platform for for the two of you do you have any advice on this project well basically it's fantastic that China is stepping up to fund more basic research whether it's in medicine or computer science you know that's going to help strengthen its universities it's going to speed up Innovation president she talks a lot about the importance of innovation so China has a lot contribute you have to be careful how you structure the government's relationship to Innovation for example in Japan they pick some very specific goals and approaches so-called F generation project and it

24:00was the wrong way to go in fact in the 1980s the US was feeling like okay Japan's better than us they know how to make things better than us they know how to plan a government AI project better than us they know how to do HD television better than us and actually that was a very good period in the US because people were kind of humbled and they looked at our sort of scattered system of many universities and no Central plan but in fact the work in the 1980s was the very best work that was laying the foundation for so many advances including the PC the internet was being done by a very distributed system that funded a bunch where put aside the Manhattan Project because that

24:41was special and it act actually worked the way they did their computer science stuff was backing dozens and dozens of different approaches and having contests like the DARPA driving challenge now they're having this pretty cool DARPA robot challenge so I would highly recommend not doing what Japan did and looking at why the US model decade in and out has has been the

25:10best I agree with what what will said that I mean it's important to create an environment that Fosters Innovation and but you wanted you wanted to let it sort of involve in aan way as opposed to of you order at a sort of a high level level of pick a a technology and decide that's that's the thing that's going to win because it may not be and you know just you know it's really you should really let let things involv but I mean but if there's High L support for Innovation and and U and and particular

25:51make sure the penalty for for for failure on is is low you don't want to fail to punitive that's one of the keys silic Valley success there many found companies very successful but in the past the company failed but they just they quickly reconstituted people left and joined other companies and it's really good at recycling and going through recursive self improvement

26:24so that it's that that's I think a really critical thing if you're going to say you want to support Innovation and new technology which is you don't know what the path is there's no math there's no light nature is unknown which means you're going to make false moves so you must it must be okay to make did you have a company before PayPal or was that your first I did yeah it's a two that was before and did you have one before bu not really I I have some ideas but

26:59it was not in the form of a company in the form of a startup yeah speaking of that I I think you you became famous way before and I started our company so you probably have the chance to know some of the last generation entrepreneurs is there anyone that you have U contact with and you have respects and and you you learned things from is there anyone like that well the you know when I started Microsoft IBM was the dominant company and they were sort of this be and then we ended up creating together with them the so-called IBM personal computer ships in 1981 so it's kind of funny to be working with them and we learned a lot from them there are some great people there like Frank

27:59kry then we worked a lot with Intel I learned a lot from Andy Grove and he was not always nice to me but that still allowed me to learn well you you were Partners right went yeah that partnership was more challenging than it might appear from the outside but but anyway Andy's was a brilliant person some of his books about management are very good Gordon Moore who did more of the chip level technolog is amazing person I worked with and he's a also a great philanthropist now you know Steve Jobs was an incredible person and part of the reason he was so incredible is he was so good at engineering great products even though he didn't know much about

28:44engineering his ability to pick people and Define what the product should be without really understanding the rules of software our chip actually make his C all the more amazing and the way he built groups and the way he he pulled things together so there were a ton of great people I learned from I may have learned more from Moren Buffett who's completely outside all the crazy computer industry stuff but his basic views about business and how you work for business actually had more influence on me than any other single person also I'd say having the Japanese for customers even though I'm saying in the 1980s they weren't structuring things well their emphasis

29:29on quality and Microsoft had about 40% of the sales coming from Japan for five early years their emphasis on quality really woke us up and changed a lot of how we did things so that was a huge evolutionary push for us so is there any any more outside of Technologies like I know Warren Buffy but he he said more I know youest right and in a new that you had contact and you learn things

30:03well now that I'm in a phase of philanthropy I'm getting to meet a lot of incredible entrepreneurs in a lot of different business areas but when I was starting Microsoft and people would say to me you're an entrepreneur I'd always say what does that mean I'm a software guy you know I didn't wake up and decide shall I do I'm I'm a generic entrepreneur and should I do barber shops or cookies or oh so I was a

30:31software person you know at age 12 I was writing software at age 13 I was writing software there was only one thing that I had a chance to be world class at which was the thing I was obsessively doing from age 11 to 17 and so it's not like I could have started a company doing anything else in the universe you know I'm I'm a softare person and then I had to learn oh God you have to hire people oh God you have to fire people that's a lot of fun you have to do budgets you have to sell things you got to go to conventions and have boots and stand there and you know fog your wees so I wanted to have software for myself that would be empowered for myself and then I had

31:14to learn okay there's contracts and many things that go along along with that so in some ways you know the drive all came from the dream about software not some dream about okay I'm going to have a company let's just pick an industry and go go get it done well it it's been some years since you were 13 and will you look back when you reflect on on your previous life and compare yours with the other successful entrepreneurs and when you look at the the younger generation do you see any difference between each generation of entrepreneurs I guess there are probably a lot of commonalities among different Generations but is there a sort of a Evolution that the Next

32:04Generation and nor are better or smarter or stronger do you see anything like that well I'm very impressed with Mark Zuckerberg I've gotten to know him well and you know I'm impressed the the current generation is actually pretty good about asking us old me because I'm old these guys you know what mistakes did you make how should things work you know our generation was in a sense it was a cleaner she the paper so every conference I would go to would be people our age there weren't older people and then you know all the analyst or Venture Capital people who we worked with all kind of grew up but it was an industry of people in their 20s and then in their 30s now it's much more

32:52age diverse you know Mark's what 32 or something and he's not any anywhere near the youngest guy he's the youngest guy with a mega mega success type company I think maybe this group is a little more mellow you know Steve was kind of crazy in a certain way and I was kind of crazy in a certain way maybe it's a little harder to be but I guess more mellow a Twi do do you see that no I don't know I I it's hard to characterize because there's there's very diverse personality some of whom I'd like to U meet but you know I have don't have a chance to you know like the guys at 10 C I don't know all that well Jack M I've gotten to know and

33:41you know people tend to be on their best behavior when they're meeting you know other people understand what it's like in the engineering meeting where somebody comes in and tells you oh this thing isn't working it's going to be late or hey it's just too damn slow do we have to build it rebuild it from Crouch that's where you would I would love to be in there and see how they deal with the various challenges cuz all these businesses get e externally internally to the point where they they have to rethink some of the assumptions they made and you know at every Point Microsoft is more successful people are saying well that's their final success and it really bugged me

34:23because when people were saying oh eventually they'll get it wrong I knew that was true I mean the companies I grew up admiring I wonder if anybody's everever heard of weighing and and digital equipment you know which are long long gone but those were the Great companies and it was probably good for me to see them fail because it meant that just the wonderful technology and the great people they had in no way protected them from when the when the marketplace turned left and they didn't notice that was happening that none of their Excellence protected them from missing a a market

35:01change I would say there are probably a lot of commonalities among those entrepreneurs I also read something about you you were quite rebellious when you were at College you only choose to audit those classes you didn't register for and you nonstop working for like 36 hours in class for 10 hours and and when you wake up you you you just keep on working do things that's the kind of saying sucessful Ur need to have well as I said

35:39I think Steve and I were in our own different ways not the same ways I didn't take as much acid as he did we were kind of crazy and I don't think that's a necessary part you know people debate you know could Steve have still been great and been nicer to people my answer is yes you know he was great I'm not taking anything away from that but you know in my case I think okay there were periodes very tough on on people probably because I was very tough on myself and you have to learn not to project that into your management style which is kind of a natural thing the harder thing was that I'd been a you

36:21know in my school I I was known as the guy who was good at math and so then I went to Harvard I was in this class with 80 people all of whom's personal positioning were they were the best person they'd ever met at MTH and so 79 of us were a fraud eventually I came in second in the class and the guy who came in first is a very successful lawyer in New York City but this idea of never attending classes that was just have some unique positioning that you I wasn't the same as all those other 80 guys it's like years and yeah I didn't expect to create companies or anything like that so it's it's really and I think it's it is it is better to approach this from a

37:06standpoint of saying rather than like you want to be an entrepreneur or you want to make money it's like what is some useful thing that you could do that you wish existed at and then try to get other people to work with you to create that thing and keep making it better and better and if you do create something useful then money your but it's but that that's just the way that the you know properly working economy kind of rewards the creation of useful goods and services and if you're creating something that you love and you think other people will love it's very it's much easier to of sacrifice the time and spend the effort and and if it doesn't work out you don't

37:49regret it yeah but have you ever s of this saying that okay I'm I'm younger than b Gates and I don't want to exactly copy him is there any difference between you know substantially some similarities growing up and you a lot of Prett much everything both said here but yeah my life kind took me in a different direction I said most of it most on software and then yeah I was like

38:29well do seem like people are really s electric car thing or space thing and so I thought i' give it a try proba that was and and I had to learn how you make hard so I didn't even know I've never seen a CNC machine or L up carbon fiber or know any of these things but but if you read books and talk to experts you will pick it up pretty quickly fact I think people self liit their ability to learn it's it's

39:16really pretty straightforward this read book and yeah and particularly particularly books I mean you can the data rate reading is mucher when when somebody's like talking I you know what's what's the output rate of speech a couple hundreds per second I mean maybe fews per second you're really going so you can do several times that

39:49reason I because the DAT too slow it's like why are they why are they reading the textbook to me this is like a story orous all right since we at B this is not a pure technology conference there are lot government officials coming here yesterday I think both of you met this and about a week ago

40:26Singapore founded funding Premier know pass away he's very well respected Statesman presy also said he made outstanding contributions to the peace and development of the Asia region I I think Chinese leaders from done to meeting and M they share something in common they believe in this kind of technocratic approach running a country or government using an engineering type of methodology this is very different from the US I think us are pretty much R my lawyers what what are the pros and cons

41:18in your eyes between these two different systems well I mean being being an engineer obviously to favor the engineering approach I think in the US we probably have too many lawyers and they shouldn't be lawers we probably have more than we should have I generally I think it's good if te and So Pro

41:52te approach yeah I think it's I I I wish there were more people in the American Congress that were proficient in in science you met a lot of Chinese government officials different generation of leaders they probably have more to say about this well I was very lucky that I got to spend have a number of dinners with Le Juan youu and he was not only just thoughtful just the very idea that he would take parts of the western system and oh this part is good this part may not apply everywhere this part he disagreed with it was kind of bold because of course the Western system was succeeding you know basically all the rich countries in the world had followed the Western system

42:42and so the idea that he thought he would do it slightly differently was a huge contribution and you know it singapor is a city state so you can do things in terms of paying government salaries and there that may not scale up but what he did was very incredible what we really want is this mix of democracy and

43:09expertise and no country has that balance right if you air on the side of democracy there's certain extreme things about the wrong person getting in power and if you kick them out and how do you get new people in you know so the Democracy side has some huge advantage ages that have helped the US

43:32quite a bit it is a little scary now when we have complex problems like how do you run a health care System efficiently why are why is the us paying so much and the there really isn't this time any elected representative you can have a good discussion about the Dynamics of that system and why it's different than other countries and how we might change that so government has to De with very complex issues and the chin government although I'd say the trend is a tiny bit away from it has had engineers and scientists in a lot of the key positions and you know willingness to look at what other countries do and also this notion that you're going to

44:11have new policy a often try it out in part of the country see if it works and then to the policy before you try and scale it up in a very broad way so the Chinese government is a student of policy more than just a a say UK Parliament US Congress where people are kind of yelling each other about you know I'm right no you're right it's not like oh let's go do an experiment I've never sat in the you know US Congress say oh let's try yours out and one state will try out mine and we'll come together and combine the best teachers that's not a typical electoral dialogue that we're having right now so it's you know it's a work in progress there are things like how you run

44:54universities where the US model you know other people should just adopt it then there's things like Health Systems and governance where they maybe should take some aspects but try try out variations so that you know we get the benefit of you know 192 countries of slightly different experience including at the subnational level so we have a very strong government they have very strong execution capabilities when it comes to infrastructure but highspeed well or

45:29highways we have massive constructions and now it's probably the largest infrastructure in terms of Transportation in the world but when you have a strong government are you concerned about Innovation there might be things that that's too strict that that that's C Innovation well I don't I think training ation is going at a very high speed I think that the government could raise how much it spends on science research even more than it is it's it's it's raising those things you know there have been areas like in the media where you have to say why the Chinese market although this is starting to change it's been behind in terms of media companies and things I

46:23think a little bit the the state probably deserves as a little bit of the reason the media sector hasn't come further along but in terms of things like how do you make energy you know the state policies are not holding back somebody you know figuring out some big invention in fact now I have a company at a nuclear power company called Terra power that really China is the most natural partner for us with the Breakthrough generation of nuclear because China's a lot like the United States was in the 1960s where the idea you want to go forward and do new things it's very clear the idea that the state's Co isn't where you want to be the us today is very careful that hey

47:05we're pretty happy with the current conditions so if somebody wants to build a new building or take some new approach there's a lot of maybe no you know there's like five levels that you go through maybe no maybe no and so it's U whereas the bias towards moving and doing new things which has a small downside but a huge upside as well I'd say in terms of breakthroughs in some areas like nuclear it's more likely to come out of China than almost any other place because of this bias towards doing big projects and in in know 1950s 1960s after was the US in the 70s it started to be Japan Korea took on that role and you know that big engineering u u bias is is is great for

47:54the world so even know you are very futuristic let's say 50 years from now how would you Invision a conference like B be done do you think we still have to come to the stage and meet face to face or there a better way for people to communicate as a comp the airport will be closer

48:27I think people still want to do conferences if only for sentimental value you know it just feels better to be in proximity to people I mean even if you had really perfect virtual reality if you put on your head here and your hatic suits and everything and kind of be there it really feels like you're there it's still the same as actually being there I think people

48:57speaking of Transportation there are lots of U bus about self-driving cars how long do you think it it's going to take for self-driving cars to become mainstream in Us in China and other part of the world sure I mean let me prise this by just saying that the with something where there's an enormous industrial base like like cars anything that happens on M is going to take time there about 2 billion Clos trucks on the road Global capacity is production capacity is about 100 million truck here so let's say tomorrow we

49:39had self-driving cars work perfectly it will still take a long time to switch up the mass industrial base take 20 years there were the same number of C Tru probably some smaller number Tru things for Economist but but it's still it still will still take take a while so so I think but to answer the question say where do I think the the first production self-driving car will exist that people can com buy it's probably well in terms of the technology working I I think we're more than 5 years away from from having cars that are way safer than VI yeah but for mainstream adoption there need to be laws and the government support that's why I said from a technology standpoint

50:25so from the point which you have the technology working then you have to The Regulators that is truly safe so You' be prly running biologically cars in Shadow mode where there's human driving and you evaluate what the what the narrow AI on the car was doing compared to the Human Action and then you able to statistically prove that the thei of the car would have been safer than than and it's going to take at least or years so from the point at which it actually works I would expect there be some delay in the regat approval on the of at least two or three years yeah I might point out to that is that it won't be monolithic that is companies like Microsoft have keep

51:16coming in all time saying we're going to do the most advanced city in the world well now it's clear if you want to design a city that assumes that drones are allowed to deliver things and you know come up with ways of of sort of on board client side collision avoidance and you want to assume you need no parking garages because these self-driving cars are around we should have a few dozen jurisdictions in the world to decide to get out there in front get the benefits and the the challenges of that just so we get you know hundreds of millions of hours of this software seeing you know the ball come into the street and all that Pioneers you know do there's some

52:00downside with it but so that's why it's good we don't have one government even in China they should pick some places and have super liberal rules so I think we're a few years away from having say a dozen jurisdiction say let's just go for it and Design This in and then in a decade we'll see the benefits as well as a few of the problems so besides the government is there any other factors or or type of people who need to be prepared about this U you know driver these cars what kind of preparation for People Like Us need need to have with the benefits are mainly when you design a city from scratch assuming that you get you get this thing we don't

52:49have that kind of luxury in most cases that's rning mega cities and how how can we smoothly transition to that kind ofet you can have a regulatory environment where things like uber are allowed to come in Uber's a form of of a driverless car it's just using sort of on demand human labor to efficiently provide that service and you know so you think about Uber for everything the way we use Capital Equipment and it's it's pretty exciting that the very Market mechanism of when should be used in a different way was very inefficient until we had this digital template and now we're seeing that the matching of I want to use an asset I have an asset has been amped up that

53:41that matching function is way less friction now than than ever before so it's changing a whole ton of things about capital goods and and labor usage so uman do you see Uber as part or competitor what's going to happen when nobody need to drive the car I think you know NE a partner or a competitor I mean we're just making cars so like we were can buy the cars other people can buy the cars and yeah and they'll gradually get more autonomous and but again I do want to emphasize that that the point I made about the industrial based switch over means this is overnight it's literally impossible because billion C

54:33million Max so should be Pani jobs the moment self cars yeah who knows what will happen in that environment Pros well one hour it's a very short time but we I think we covered a lot we talked about history we talked about future and I think in times history can tell us what the future will look like and I sincerely hope that U Our Generation can learn from the history and make contributions to

55:26the the next few Generations so that they can benefit from our effort and with that I thank you all those for coming through this session thank you [Applause]

资源与出处

相关视频

同频道、同主题或相邻年份