大会演讲·中文精读

对话Tesla CEO马斯克:2017迪拜世界政府峰会

A Conversation With Elon Musk, CEO of Tesla Inc At World Government Summit 2017 Dubai

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

2017年2月,迪拜世界政府峰会。台下坐着上百个国家的政府代表,主持人把马斯克和牛顿、爱因斯坦、莱特兄弟并列介绍了一遍。

然后他开口聊的第一件事,是电子游戏的画质。

从四十年前那两个在屏幕上弹来弹去的方块,一路推到虚拟现实,再推到“我们也许活在一个模拟出来的宇宙里”。在场的人大概没料到这个开场。

我把这场对话完整读了一遍。关于 AI、自动驾驶、隧道、火星的那些判断,放到今天回看,有的已经兑现,有的还在路上,还有几条当年说出来没人当真。

一切从“人生意义是什么”开始

马斯克说,他十几岁时有过一段存在主义危机。他想搞清楚人为什么活着,读了一堆哲学书,越读越困惑。

最后他自己得出一个结论:真正重要的是“问对问题”,而人类意识的范围和规模越大,能问出的问题就越好。

这个想法几乎解释了他后来做的所有事。要让意识延续下去,得先解决两件底层的事。一是能源和交通的可持续,不能靠烧化石燃料把自己烧死。二是成为跨星球的物种,把意识备份到别的地方。

他把后者比作人寿保险,一份替“人类意识”这个整体买的保险。说完他又补了一句,官方的解释到这儿就够了,真正驱动他的东西在别处。

一幅图景是我们永远被困在地球上,直到某天出大事;另一幅是我们散落在很多星球上,甚至铺满整个太阳系。第二幅图景令人兴奋得多,也鼓舞人心得多。人早上起床需要有理由。生活不能只是解决一个又一个问题,那样的话意义在哪里。(马斯克)

这段话里藏着一个容易被忽略的判断:提出问题这件事,才是文明最核心的产出。火箭、电动车、太阳能,都是为了让“提问”这个动作能一直持续下去。

游戏画质那条曲线

主持人问他,人不该有灵魂、有信仰吗?马斯克绕了个弯回答。

他说小时候玩过的最贵的游戏,屏幕上就两个方块加一个点,来回弹。当时觉得挺好玩。现在的游戏已经接近照片级真实,几百万人同时在线,虚拟现实还在往前推。

然后他做了一件很马斯克的事:把这条曲线往外延伸。

哪怕每年只进步 0.1%,推得足够远,游戏和现实的差距就会缩到肉眼分辨不出来。技术是单向的,只会变好,不会倒退。

那问题就来了:我们怎么知道,这件事在过去没有发生过?(马斯克)

他说得挺轻松,主持人在旁边笑。这个论证的锋利之处在于,它不要求你相信,只要求你承认“无法证伪”。

四十年的画质进步是已经发生的事实,往外推是纯逻辑。你不需要接受结论,你只需要找出这条推理链断在哪一环。大多数人找不出来。

预测未来的正确姿势

主持人问他,技术变化的速度到底在变快还是变慢。马斯克说这类事情特别难预测,然后给了两个数据点。

第一次可控动力飞行是 1903 年,莱特兄弟,飞了十几秒。66 年后,人类把人送上了月球。

从“人类第一次离开地面”到“人类在另一个天体上走路”,中间只隔了一代人多一点的寿命。在场的人听到这个对比,鼓了掌。

第二个数据点跟价格有关。2017 年,一百美元的设备就能让你和地球另一端的人视频通话,几乎是免费的即时视觉通信。任何问题,搜一下,答案立刻出现。

他提醒台下的人:这些能力在几十年前会被当成魔法。

今天做的任何关于未来的预测,我敢肯定是错的。方向上也许对,具体细节一定错。(马斯克)

然后他给了个 50 年的方向:人类在火星上有据点,频繁往返太阳系各处,同时开始为飞往邻近恒星系做准备。他说这些在 50 年内都有可能实现。

2017 年听这话像狂想。今天再看,星舰还在做在轨加注的试验,时间点往后推了,方向没变。

十年后,开车会像按电梯按钮

他给了一个具体到年份的判断:十年内,一辆不具备完全自动驾驶能力的车会变得很不寻常。

理由跟算法突飞猛进关系不大,主要是硬件已经铺好了。传感器成本在下降,车载算力足以让车开得比人更安全。剩下的主要是软件迭代。

他举了电梯的例子。以前电梯有专职操作员,你走进去,告诉操作员去几楼,他帮你按。后来自动化了,你自己按按钮,一开始大家还不太习惯,现在没人觉得这是个问题。全自动驾驶会走同一条路。

2016 年 10 月,特斯拉开始给新车装全套自动驾驶硬件,当时被不少人嘲笑“装了一堆用不上的传感器”。马斯克说其他厂商都会跟上。后来确实都跟上了,只是节奏各不相同。

真正的风险在“广义”

主持人问 AI。马斯克把它切成两块。

狭义 AI,比如车辆自动驾驶,目标是完成一个特定任务。这类东西可控,风险有限。

另一类是他真正担心的:智能水平接近甚至超过地球上最聪明的人的通用 AI。他的原话是,那是一种危险的情况。

他举的例子很具体。假设 20 年内会出现超级智能,在这之前,我们身边已经有一群比我们聪明得多的“东西”,我们对它们的行为规则却还没想清楚。谁来踩刹车?

他给政府的建议也很直白:盯紧 AI 的研究进展,确保科学家不会因为沉浸在技术突破里而忽略后果。科研共同体有一种惯性,做着做着就忘了自己在造什么。

这段话当年听是提醒,现在是常识。2023 年那份要求暂停超大模型训练的公开信,签名的人里就有他。

关于外星人

主持人抛了个宗教性很强的问题:这片土地上的人相信人类不是从这来的,你怎么看?

马斯克的回答分两层。

第一层,如果宇宙里真的存在超级智能文明,它们大概率已经在观察我们了,只是我们不知道。

第二层是个尺度问题。一个掌握了光速 10% 到 20% 航行能力的高级文明,把整个银河系铺满,大概只需要一千万年,最多四千万年。放在宇宙一百三十多亿年的历史上,这点时间几乎等于零。

结论是:我们至今没看到任何迹象,这件事本身说明了一些东西。至于到底是什么,他说他也不知道。

这就是费米悖论的一个通俗版本。他在几百个政府官员面前承认自己没答案,这一点比答案本身更有意思。

为什么要死在火星

这个问题问得有点黑色幽默。主持人提到他公开说过想死在火星上,问他为什么。

马斯克的回答挺实诚。

我不想死。但如果一定要选一个地方死,为什么不选火星?我们生在地球,为什么不死在火星上。(马斯克)

现场笑了。但这句话的分量在后面:一个愿意“死在那儿”的承诺,比任何商业计划书都更能把人聚起来。

去火星这件事,短期内没有任何财务回报,风险极高,参与者要接受可能回不来。这种情况下,创始人自己先把命押上去,是唯一能让团队相信“这事真的要做”的方式。

堵车的解法在脚下

他说交通拥堵的解法是三层隧道网络。

注意是三层,不是一个平面上的几条线。城市拥堵的根源在于所有建筑的地面出入口都挤在同一个平面,所有人上下班的时间又几乎一样。

地下没有这个限制。他给了个很好记的对比:人类挖过的最深的矿井,比建过的最高的楼还要深。往下挖的空间比往上盖的空间大得多。

20 层、30 层、40 层、50 层的隧道网络,理论上可以彻底解决任何一座城市的拥堵。

难点在成本、速度和安全,不在工程原理。盾构机太慢,一米一米地啃,造价高得吓人。他的思路是反过来改机器本身,把掘进速度提上去,把成本压下来,把安全性做到可接受。技术曲线一旦拐过来,这件事就成立。

后来他成立了 The Boring Company。2017 年说这话时,很多人当笑话听。

关于中国和这个地区

主持人翻旧账:2015 年 6 月 4 日我见过你,问你能不能来这边发展,你说你忙着中国的事,可能短期内不会来。一年半过去了,现在有变化吗?

马斯克承认,早几年在中国确实遇到一些硬问题:充电网络、售后服务、各种基础设施。但他给出的判断是,这些问题现在解决得不错,中国业务正在跑起来。

然后他说,现在是认真进入这个地区市场的好时机,从迪拜开始。

这段话在 2017 年是给当地政府的一个善意表态。放到今天,阿联酋和沙特的电动车渗透率、特斯拉在中东的布局,已经能看出这条线索走到了哪儿。

全自动驾驶会砸掉最大的饭碗,但时间表是 20 年

这一段数字密度最高,也最容易被误读。

他先说结论:全自动驾驶对就业的冲击会很快到来,可能是最大的一次职业替代。理由简单,驾驶在各种形态下,可能是人类最大的单一雇主。

然后他立刻做了个自我修正,把“快”这个字拆开。

全球有超过 20 亿辆车,正在逼近 25 亿辆。每年新造车的产能大约一亿辆。一辆车的寿命大概 20 到 25 年。

这三个数字凑在一起,结论就清楚了:技术成熟的那一天,和全社会大规模失业的那一天之间,隔着很长一段时间。因为要把路上跑的二十多亿辆车换一遍,本身就得二十年。

真正的冲击会在差不多 20 年里展开,涉及劳动力的 12% 到 15%。(马斯克)

20 年听起来很长,放在社会结构里其实极短。一个现在 40 岁、靠开车吃饭的人,60 岁那年会遇到一个不再需要他的市场,而他的再就业窗口在前十年就已经关上了。

给政府的建议:盯 AI,准备电

主持人问他在座一百多个国家的政府代表:你建议他们现在做什么?

他给了两条。

第一条关于 AI。政府要保持对 AI 研究的关注,确保公共安全不被技术冒进牺牲掉。他特别提到科研共同体的自我驱动问题,科学家在做技术突破的时候,未必会同步意识到这个突破意味着什么。

第二条关于电。

他说全球能源使用里,电力大约占三分之一,交通(几乎全是化石燃料)大约也占三分之一。如果交通全面转向电动,电力需求大概要翻三倍。

这个数字对政策制定者的含义很直接:电动车的普及速度取决于电网的扩张速度,充电基础设施是前置条件,不是配套工程。谁能提前把发电和配电能力做上去,谁就能接住这波转换。

他自己说,这个转换大概要 30 到 40 年,比自动驾驶慢得多。

全民基本收入不是最难的部分

被问到自动化导致的失业,马斯克说得非常直接:最终大概只能搞某种形式的全民基本收入(universal basic income),没有别的选择。

理由是产出。全面自动化之后,商品和服务的产出会极高,几乎所有的东西都会变得很便宜,那是物质层面的丰裕。

然后他把话题拐到了更难的地方。

更难的问题是,人怎么获得意义。很多人是从工作里获得意义的,如果没有工作,你怎么找回意义?你会不会觉得自己没用?这个问题难得多。(马斯克)

他说到这儿没有给答案。一个能造火箭的人在这一题上也没有答案,这件事本身就值得记住。

政策讨论里,全民基本收入的金额、发放方式、财政来源被反复算过,但“人为什么早上要起床”这一层,几乎没人真的回答过。

我们已经半机械化,缺的是带宽

他给了一个方向:生物智能和机器智能的融合。

他说这话不是科幻。现在的人已经是半个赛博格了。手机、电脑、各种应用,构成了你的“数字第三层”。你可以在几秒钟内调出全人类的知识库,这在几十年前不可想象。

他指了个细节:一个人的数字层在物理死亡之后还活着。你的邮件、照片、社交媒体账号,还留在那儿,继续存在,继续被检索。

他把融合看成必然会发生的事,唯一的问题是融得好不好、多快。他关心的指标只有一个:带宽。

他的说法是,计算机之间的通信是万亿比特每秒的量级。而人往外输出信息的通道极其狭窄,键盘敲字,拇指在屏幕上戳,速率大概只有每秒十比特到一百比特这个量级。大脑里想清楚了,输出的速度慢得像用吸管倒水。

后来他把这个想法做成了 Neuralink。逻辑在这段对话里已经讲完:提高人机之间的带宽,让人类在智能爆炸的时代还留在牌桌上,顺带也许能解决 AI 的控制问题。

火箭复用的秘密就是“飞机为什么便宜”

主持人问他为什么非要啃火箭复用这块硬骨头。

他用了飞机做类比。如果每次飞完一架飞机就把它扔了,机票价格会是天文数字,航空业根本不会存在。飞机之所以便宜,是因为同一架飞机能飞几千上万次,制造成本被摊薄到每次飞行里几乎可以忽略。

火箭是一样的。一枚火箭六千万美元,燃料只占其中很小一部分。如果能复用一千次,分摊到每次的资本成本就是六万美元这个量级。再配上单次运载很多人,单位成本就能逼近航空的水平。

然后他说了难点在哪。

地球的引力相当大,把火箭做成可复用,比把飞机做成可复用难得多。这需要最好的材料、最好的设计,把每一个环节都做到极限。任何一处差一点,整件事就不成立。

十年后回看,猎鹰 9 号的一级回收已经成了常规操作,单枚助推器复飞十几次。他当年说的“很难,但可以做”,被验证了两遍。

面试只问一个问题

主持人问他怎么招人。他说他面人的时候基本只问一件事:讲讲你的经历,以及你一路做过的那些决定,为什么那么决定。

接着是第二个问题:讲几件你解决过的最难的问题,以及你是怎么解决的。

他解释了为什么这个问题有效。

真正解决过难题的人,能把过程讲到很细的颗粒度。他们记得那些卡住的地方,记得当时试过什么、为什么失败、后来怎么绕过去。假装解决过的人,通常只能讲一层,再往下一问就露馅了。

这套方法验证的是“你亲手做过”,“你知道有这回事”过不了关。

最大的挑战是维持反馈回路

主持人问他最难的是什么。他答的有点出乎意料:怎么把时间分配好。

再往下一层,他说更难的是确保自己拿到的是正确的反馈,并且长期维持这个回路。因为位置越高,愿意告诉你真话的人越少。人们会倾向于说你想听的话,久了之后你会活在一个失真的世界里。

这一条和后面给年轻人的建议,其实是同一件事的两种说法。

给年轻人的建议

主持人问他对在场想成为“下一个马斯克”的年轻人有什么话说。

他先泼了盆冷水:大概不该想着成为我。

然后给了两条实际的东西。

第一条,学物理的思维方式。他强调的是思维方式,公式不在此列。物理训练的核心是:先假设自己是错的,然后把目标定为“比昨天少错一点”。你要主动去攻击自己的结论,去找能推翻它的证据。

第二条,警惕一厢情愿。他承认自己也会犯这个错。人一旦希望某件事是真的,就会忽略那些说明它不是真的证据。

这个陷阱极难避开,我自己也经常掉进去。(马斯克)

他还补了一个具体做法:主动去要批评意见,尤其是从朋友那里要。因为朋友在乎你的感受,他们不会主动告诉你哪里错了。你得明确说:我真的想知道。

这场对话里最值得拿走的东西

整场对话,马斯克说的所有判断都遵守同一个模板:先给一个反常识的结论,再给支撑它的数字,最后说清楚这个判断会在什么条件下失效。

游戏画质的推演,前提是“技术进步是单向的”,这个前提可以被质疑。自动驾驶十年普及,前提是“硬件成本继续下降”,这个前提也可以被质疑。

但他把前提摊开了放在桌面上。你不同意他的结论,可以;你不同意他的前提,也可以。他给了你怼他的接口。

大部分关于未来的讨论里,人们只给结论,不给前提。那种讨论听完了什么也留不下。

另一件事是他反复提到的“反馈回路”。位置越高越听不到真话,越听不到真话越容易做错判断,这个正反馈循环一旦启动,很难停下来。他把它称作自己最大的挑战,比造火箭还难。

这句话从一个同时管着两家高难度公司的人嘴里说出来,值得多想一会儿。

  • 把你对未来的判断写成带日期的文档。今天写下你认为 2030 年会怎样,三年后翻出来对照。没有记录,你永远不会知道自己错在哪、为什么错。
  • 面试时把“讲过程”变成标准动作。别问“你负责什么”,问“你解决过最难的问题是什么,怎么解决的,中间卡在哪”。回答只能停在一层的人,多半没真正动过手。
  • 主动设计你的真话通道。每季度找三个敢说话的人,明确问一句“我最近哪里做错了”。朋友默认不会主动说,你不开口,这个信息就永远拿不到。
  • 如果你在做基础设施相关的决策,把电网当成前置条件。电力需求翻三倍不是终点时的账单,是起点时的约束。充电网络的扩容速度,会直接决定电动车的实际上限。
  • 给自己设一条“证伪清单”。把你最相信的三个判断列出来,分别写下“什么证据会让我改变想法”。写不出来的那一条,多半是一厢情愿。
  • 看待自动化替代,盯住置换周期而不是技术发布日。技术成熟和社会结构变化之间隔着十几年,这个窗口是给人做准备的,不是给人安慰的。
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)。

0:01foreign [Applause] [Music] foreign and foreign [Music] foreign

2:01[Music] foreign here [Music] because [Music] yeah [Music] [Music]

3:40businessman it is foreign [Music] I'm still alive foreign [Music] foreign

5:33[Music] economics foreign [Music] welcome strategies exercise [Music]

7:12technology foreign [Music] foreign Shane I'm going to strategies [Music]

8:55foreign [Music] driverless Transportation by 2030. how will we get there [Music] future for everyone and be driven by no one introducing the new vz10 parts of our plan was to testify those Transit technology an electric Shuffle dedicated to Smart Mobility designed to take people from transportation to probably a short distances at predefined roots with no residence [Music]

10:56a reality energy isn't isn't aggression of Direction [Music] and transports Authority [Applause] is

11:54foreign [Music] foreign foreign [Music] foreign [Music] foreign

14:57[Music] foreign [Music] foreign [Music] foreign [Music] reporting sounds

17:11six point s [Music] oh yeah [Music] foreign [Music] I'll be asking results foreign

18:49[Music] foreign [Music] come on foreign [Music] foreign . anymore

21:58foreign [Music] foreign [Music] foreign [Music] similar disease [Applause]

25:24[Music] foreign [Music] Airlines is New York tonight seems to distinguished guests now there's a conversation with Elon Musk CEO of Tessa corporate the founder of SpaceX moderated by his Excellency minister of Canada's Affairs and Future German of the world governors [Applause]

27:37[Music] challenge to move through one of the most in my opinion in life we've seen within this part of the world great people like me who invented either nobody Newton any more the right brother Albert Einstein and Elon Musk [Applause] see you are impression you want to go to places that nobody did

28:39you already lifting acceptance from the rocket industry to space it to the car industry with Tesla Muslim mission one you do whatever we do thank you for having me sorry we're here and it's really good fantastic I'm really encouraged anyone who Transit into business and those are the motivations against the the Lord with the long version of the explanation but

29:29essentially the when I was a kid I was wondering kind of what's the meaning of life like why are we here and I came to conclusion that what really matters is trying to understand the right questions to ask and the more that we can increase the scope and scale of human consciousness the better we're able to ask those questions and so so I think there's certain things that are necessary to ensure that the future is good and some of those things are in long term having long-term sustainable transportants

30:19and to be a spacecraft civilization songs and species I think the biggest size is important for the long-term disciple of human and that's one reason kind of like life insurance but then the part that I find personally most motivating is that is that creates a sense of adventure and it makes people excited about the future but if you consider three features one where we are forever confined to it and So eventually something terrible happens or another picture of where we are out there on many planets maybe on the solar system I think the second version is incredibly exciting with an inspiring and they need to be reasons to get up in the morning you know life can't just be

31:24about solving problems otherwise what's the point there's going to be things that people find inspiring I mean I help you before speaking this is step three the zip is not anything is life or illness as I get older I find that question to be maybe more and more confusing or probably or uncertain

32:05they particularly when you see the advancement of something like video games you know like say 40 years ago you know video games the most expensive video game would be like like home where you have like two rectangles and a dog you know like batting a back and forth oh yeah exactly it sort of takes you a little bit but yeah and that was like wow that was a pretty fun game at the time but now as you can see the video that's photorealistic almost very realistic and millions of people playing simultaneously and you see what it has been going with virtual reality and if you extrapolate that out into the future with any rate of purpose at all like even

32:560.1 percent or something like that a year ago that eventually those games will be instinctual in reality it'll be so realistic you will not be able to tell the difference between that game and reality as you know and then it seems like well how do we know that that doesn't happen in the past El

33:24I mean if anything is possible I mean there's a lot I mean yeah give me like instantly accelerated to something isn't it I mean if our life see in the past 100 view like the Constitution yeah in the past 20 it's very famous like faster or faster is it more slow so when you push in the street

33:57how to see well I think it's one of those things that it's quite difficult to predict we think of say I mean the first controlled powered flight was 9303 with the right Focus and then 66 years later we put the first people on the moon foreign [Music]

34:39but today with a hundred dollar advice you can you can video conference with anyone in the world on the other side of the world for energy I'm a Wi-Fi connection you know just before it's basically free to have an instant visual communication with anyone or even with millions of people in social media communicate to one of these people's similar names so and you can Google something and ask any question that's like an article that would assume that you've been asking us for any question again the instant response we've been incredibly difficult to break these things even the relatively recent fact so I think the one thing I think would be quite certain of

35:29is that any predictions were made today for the board the future will be like in years we'll be wrong that's for sure I think directionally I'll tell you what I thought as opposed to maybe what it will be if this is the same spiritual thinking and I hope we are out there on Mars and maybe

35:58we're traveling frequently throughout the solar system perhaps preparing for Missions to nearby star systems and I think all of this is possible within 50 years and we've got to get very excited to do that and I think what we'll see autonomy and the artificial intelligence Advanced tremendously I think that's actually quite an internal my guess is and probably 10 years it will be very unusual for

36:37cars to people that are not fully autonomous 10 years yeah I think almost all cars

36:53have the sensor system necessary for full economy probably enough compute power to be safer than a person so it's mostly just the question of helping the software uploading the software and it turns out that the computer power is limited we can easily avoid the computer and so that's all we'll test those blood since October of last year and other manufacturers will follow and do the same thing so educational car will be like getting an outlet you just tell it where you want to go and it takes you there with extreme levels of safety but then we won't be normal like for elevators they used to be elevator operators again and when you die from your lever now I understand

37:43if you press the button and you think about it so economy will be widespread the you know any one of the most Trump like questions is artificial intelligence I don't think narrow AI like vehicle autonomy I remember narrow AI class if it's narrowly trying to achieve a second but deep Automotive intelligence or personal smaller the challenges where they have ai that is much not much slower than the smartest people on earth it is dangerous situation

38:30so I mean there's two of you one view is talkative with a few minutes I think it's both it's it's like what we figured is Imagine we're going to be possibility imagine you're very confident that we're going to be resilient by super intelligent aliens in let's say 10 years or 20 years super intelligent so in 18 within 20 years yeah well visual super intelligence will be like in England

39:21was other intelligent like outside this is probably but I think this is this is one of the great questions in physics and philosophy is a world area maybe there among us I don't know [Applause] [Music] [Applause] hey we are 80 I mean if you look at this part of the world they believe that human beings are not from it they came from somewhere else

40:05came from somewhere else where so in a way human being is one do you think we may contact with him is an influence of the use well that's a really tough one to say if there are super intelligent aliens out there they're probably already observing them and we just and any advanced gaming civilization that that was at all interested in 10 or 20 percent of the speed of light you could populate the entire galaxy or let's say 10 million years maybe 40 million is Max this is nothing

41:08once you said that you want to die the most why I don't because I don't want to die but it's like if I mean the world kind of lives today and if you're going to pick some place to buy then why not once you know if we're born on Earth why not dire come on [Music]

41:52sometimes compared he tweeted that you are building a family under Washington DC why what is it what's the secret one yeah exactly too stupid the well I think this Escape sounds a little I mean it seems like so much I've been saying this for many years now but I think that the solution to other congestion is a network of tunnels other

42:35cities and you I want to say I don't mean a 2d Planet funnels I mean tunnels that go make levels deep so you can always go deeper than you you can go up like the deepest minds are taller than the tallest building so you can have tunnels 20 30 40 50 levels as many levels and so given that you can overcome the congestion situation in in any city in the world the challenge is to just figuring out how to evolve Tylenol quickly at this low cost and with high safety so if kind of Technology can be improved to the point where you can build funnels fast cheap and safe then that would

43:32completely converted any traffic situations in the cities and so I think that's why I think it's an important technology and one of these Washington DC LA and most of them are the major American cities most major cities in the world suffer from severe traffic issues and unfortunately because you've got these buildings which are you're full of eggs

43:57at one level and then people generally only go in and out of those buildings at the exact same time let's come back to UAE anymore first time it was 4th of June 2015 at the office is 86. and I asked you how the person anyway and your answer was I'm busy in China maybe not in the near future yeah almost a year and a half later we have seen plan for Spotify why not well I think actually thanks very reasonably welcome China so where some initial challenges clearly I'm charging and service infrastructure and various other things but now it's actually going really well and then so the time is going to be good to really make a significant debut in this

45:03region starting in design a new opinion the news that a good thing that would come to Technologies technology the way we do business the most airtime impact of course technology standpoint is the final of the cars a fully self-driving cars I said that's going to be that's going to happen much faster than people realize so and then it's going to be a great convenience to be you know kind of as far but there are many people through his jobs it is to drive so if in fact I think it might be the single largest employer of people it is

46:01driving in various forms and so that we need to figure out murals for or you know what those you need to do but it will be very disruptive and very quickly I should characterize what I mean by by quick because the there are there are over two billion vehicles in the world approaching in fact approaching two and a half million cars and trucks the total new vehicle production capacity is about 100 million

46:42which which makes sense because the life of a car or truck reports finally distracted about 20 25 so so the point which we see full autonomy appear will not be the point at which there is massive societal of people because it will take a long time to make enough autonomous vehicles to be to disrupt

47:07employment so that that disruption I'm talking about will take place over about 20 years there's still 20 years is a short period of time to have I think something like 12 to 15 in the workforce the unemployment this is for 139 governments here if you want to advise your own official ability for the future what is breathing or be advised in human

47:50would be to really pay close attention to the development of artificial intelligence I think this is we need to just be very careful in how we adopt artificial intelligence and to make sure that researchers don't get carried away because sometimes what happens is a scientific commit so it grows when they work they don't necessarily realize the ramifications of what they're doing so I think it's important for Public Safety that we you know governments keep a co-signed on artificial intelligence and make sure that it does not represent that Foundation

48:35secondly I would say we do need to think about transport ation and business the movement towards electric vehicles sustainable transform that I think that's going to be you know good for for many reasons but okay that's something that happens immediately that doesn't happen slower than the self-driving vehicles so that's probably something that happens over 30 or 40 years if you do electric vehicles

49:18so thinking about that in the context the demand equal electricity will increase dramatically So currently in terms of total energy usage in the world it's about concurred electricity about one good transport about 138 so at about time that will transition to almost all but not all but predominantly electricity which means which means for electricity will probably tripled so it's going to be very important to think about how do you make so much more electricity and

50:05I think they have an easy job that's that's it there's no one coming in Florida no I did I think really at least these things a little bit but what's new about mass unemployment challenge and I think obviously you will have to have some kind of universal based income I don't think you'll be any choice

50:40of you yeah because there is no job the machine or boss is taking over there will be fewer fewer jobs that a robot cannot do better these these are not things that I think that I wish would happen is something similar things that I think probably will happen and since my system is corrected that probably will happen then we need to say what are we going to do about it I think some kind of anniversary that now the output the output of goods and services will be extremely high absolute automation they will they will come abundance there will be almost everything will get very cheap

51:37I think I think we'll sign up during the universal escape from this can be necessary the the harder challenge much harder challenge is how do people then have meaning a lot of people they derive their meaning from their employment so if you don't have how do you voice it again you have meaning do you feel useless these are much that's a much harder problem to deal with and then how do you ensure that the future is ending the future that we want that we survive

52:22well I mean I do think that there's a potential path here which is but having some sort of Mojo with biological intelligence and machine intelligence to some degree we are already a cyborg you think of like the digital tools that you have your phone and your computer the applications earlier you can ask the question and instantly can answer from Google other things and and so you already have a digital tertiary tertiary layer I say tertiary because the computer system kind of the and

53:25itself the so you already had that and do anything if somebody dies the digital goes around you know all of your emails and the pictures that they're going to do with social media that's still it that I so over time I think we'll probably see a closer version of biological intelligence condition s and it's mostly about the bandwidth the speed of the connection between your brain and your digital the digital extension of yourself particularly and upward if anything is getting worse you know you said like keyboards that you teach a lot now you do most important on phone and that's a very slow my computer can communicate at a trillion bits per second but

54:27your phone can maybe do you attendance per second or 100 in general I think will be something that helps achieve some meiosis and between human and machine intelligence and maybe solves the control problem and the useful is the problem

55:07your idea is so huge you want to go to space you decide to go space you did it you decided that you were that you wanna Nan you rock it back okay yeah and you are always pushing by Humanity why I think that what's what technology solution is necessary in order to achieve a particular goal and then try to make as much progress in that direction as possible so in the case of space fight the critical breakthrough is instrument specified is rapid and complete the disability of others

56:10foreign so but because you can reuse the aircraft and develop the times becomes much more affordable and the same is true of rockets so our rocket costs 60 million dollars shortly so the account will cost you can use one is is 15 0 but if the capital cost if you can use a thousand times is 60 000

57:07so then if you can carry a lot of people per flight then you can get the cost to specify to be something not far from the classic airplane it's extremely fundamental because Earth's gravity world is quite the key fairly High Gravity the difficulty of making a rocket reusable is much greater than the difficulty of making an aircraft useful

57:44Advanced Materials most advanced design techniques and you get everything just right they don't complement that you can do a reasonable impossible

58:14well it tends to be you've got to feel more than anything else so when I interviewed somebody it might be a question was always the same it's just awesome I said tell me the story of God and and the decisions that you made along the way and want to make them and then and also tell me about some of the most difficult problems you worked on and how you solved them and that that question I think is very important because the people that really solve the problem they know exactly how they sell then there's a little details and the people that pretended to solve the problem I think maybe go to one level and then they could stop what was your biggest challenge

59:14Christian July how to spend the time one of these challenges I think is making sure you have correct your feedback with and then maintaining that productive feedback loop over time even when people want to tell you exactly what you want to do that's very cool times over I'll ask you just one last question and the word government Summit we had

1:00:11so many people from so many young people from here he had an advice to the young people who already want to be like Elon Musk to save us I think that's probably I shouldn't want to be I think it sounds better

1:00:42it's definitely it could be worth for sure but it's I'm not sure I'm not sure okay I think advice of you if you want to make progress in things I think that but the best analytical framework for saying the pictures I recommend studying the the the thinking process around things like not just not the equations the equations certainly not helpful but the way of thinking in physics is that's the best framework for understanding is a primary field

1:01:37and and you know what's taking the position that you know some people be wrong and your goal is to be less strong over time one of the biggest mistakes people remember me and I'm guilty of it too is wishful thinking and I'd like do you want something to be true even if it is improved and so you ignore the things that if you grow the real truth because of what you want to be good this is a very difficult trap to avoid I'd like to say one that I find myself in having problems with but if you just take that approach of your voice to Century wrong and your goal is to feeling strong and solicit critical feedback particularly from Friends like friends particularly friends if

1:02:30somebody loves you they want the recipe they don't want to tell you about them so you have to ask them okay I said really I really do want to know thank you very much and as we wait for you to have a legend who's creating the future for Humanity to share the story his idea his vision his challenges and his hope for them thank you [Applause] [Music]

资源与出处

相关视频

同频道、同主题或相邻年份