如果一定要挑个地方死,我选火星。但别是撞上去那种死法。(马斯克)
2017年2月,迪拜,世界政府峰会。台下坐着139个国家的政府代表,台上主持人穆罕默德·阿尔·格加维(Mohammad Al Gergawi)用这句话开场,问马斯克为什么。
马斯克先做了澄清:他并不想死在火星。人总有一死,如果可以选择死在哪里,火星听起来比地球刺激一点。
这个回答把整场对话的调子定住了。接下来的一个小时,他从人类为什么必须成为多行星物种,聊到我们可能活在一场电子游戏里,再到地下隧道、自动驾驶、全民基本收入,最后落到给年轻人的建议上。中间他两次承认"这个我也想不明白"。
下面把这场对话里有信息量的部分拆开讲。
他做这些事的源头,是小时候一个想不明白的问题
格加维的第一个问题:你的人生使命是什么,你为什么做这些事。
马斯克说他小时候反复琢磨同一个问题:生命的意义是什么,我们为什么在这里。想来想去,他得出一个结论:真正重要的是搞清楚该问哪些问题。人类意识的范围和规模越大,能问出的问题就越好。
这个逻辑一路推下去,落到几件他认为必须完成的事上:长期的可持续交通、可持续能源、成为太空探索文明、把人类送出去变成多行星物种。
为什么非要多行星?他给的理由很朴素:给生命买一份保险。
这是集体的生命保险,是我们所知的生命的保险。(马斯克)
但他紧接着说,保险不是最让他兴奋的部分。真正让他愿意每天干下去的,是另一层东西。
它创造了一种冒险感,让人们对未来感到兴奋。(马斯克)
他拿两种未来做对比。一种是人类永远困在地球上,直到某天出了大事。另一种是我们散落在很多星球上,甚至走出太阳系。
后者听起来更不靠谱,也更不划算。但马斯克说,人总得有早上起床的理由。
生活不能只是解决问题,否则意义在哪?总得有些让人觉得鼓舞、让活着值得的事情。(马斯克)
这段话值得停一下。大部分人理解马斯克,是从"人类灭绝风险"这个角度切进去的,好像他是一个被恐惧驱动的人。他自己给的答案恰好相反:恐惧是备份计划,冒险感才是主引擎。
我们有没有可能活在一场游戏里
格加维问了个更玄的问题:对你来说生命是什么,是梦吗,是真实的吗,值不值一百万?
马斯克说,这个问题他年纪越大越觉得糊涂。
他从电子游戏讲起。四十年前,最先进的游戏是《Pong》,两个矩形加一个点,来回弹。当时觉得挺好玩,他自己也玩过,格加维也玩过,台上两个人因为这款游戏暴露了年龄。今天的游戏已经接近照片级真实,几百万人同时在线。虚拟现实、增强现实还在往前推。
然后他做了一次外推:不管进步速度多慢,哪怕每年只有0.1%,足够长的时间之后,游戏和现实就没法区分了。
如果游戏真实到你分不出它和现实的差别,那我们怎么知道,这件事在过去没有发生过?我们是不是自己就在某个游戏里?(马斯克)
格加维的反应是:一切皆有可能。
马斯克没有往下推,只是把这个问题摆在台上。这个说法后来被叫做"模拟假说",2017年讲的时候还没那么流行。他的论证方式很物理:不谈哲学立场,只做量级估算。
今天对五十年后的任何预测,都是错的
格加维把话题拉回现实:二十年、三十年、五十年后,教育、交通会变成什么样?
马斯克先泼了盆冷水。
他举了莱特兄弟。1903年完成第一次可控动力飞行,66年后人类登上月球。如果1900年问一个普通人登月的概率,对方会觉得你疯了。跟他讲互联网,他根本听不懂你在说什么。
今天的情况是,一个一百美元的设备就能和世界任何角落的人视频通话,有Wi-Fi就基本免费。社交媒体让你同时对上百万人讲话。搜索引擎像一个随叫随到的神谕,什么问题都能立刻给答案。
这些事情在过去极难预测,哪怕是在并不遥远的过去。(马斯克)
所以他给出一个很确定的结论:今天对五十年后做的任何预测,都会是错的。这一点他有把握。
能做的是说希望看到什么,而不是预测什么会出现。他希望五十年后人类在火星上,也许在木星,在太阳系里频繁往返,开始为去邻近恒星系统做准备。他也承认这可能是厢愿。
十年后,造一台不能自己开的车会显得很奇怪
话题转到自动驾驶,马斯克给了一个具体的时间表。
他的判断是,大概十年后,不是完全自动驾驶的车会变得很少见。特斯拉当时生产的车已经装了完全自动驾驶需要的传感器套件,算力大概也够,剩下的是软件开发和推送。如果后面发现算力不够,换一台计算机就行。
他还提了一个时间点:从2016年10月之后生产的特斯拉,硬件上就都具备了。
坐进车里会像坐进电梯。你告诉它去哪,它送你到那,安全性极高。这会变成理所当然的事。(马斯克)
他补了个类比:电梯以前是有操作员的,一个专门的人在里面推拉杆。现在你进去按个按钮,没人觉得这有什么了不起。
这个细节说明了他对自动驾驶的心理预期:这东西最后会变成一件无聊到没人再提的基础设施。酷炫只是一阵子的事。
他最担心的东西,是那个"外星人"
格加维问,人工智能到底是在帮人类还是在威胁人类。马斯克说,两个说法都对。
他先做了一个区分。车辆自动驾驶属于"窄人工智能",只针对一个特定功能。真正让他不安的是深度人工智能,也就是通用人工智能,那种比地球上最聪明的人还聪明的AI。这种局面在他看来很危险。
他用了外星人这个比喻。
假设你非常确定,十年或者最多二十年之内,会有超级智能的外星人来访。数字超级智能就会像那样一个外星人。(马斯克)
这个类比的作用是:人们对"外星人入侵"有本能的警觉,对"服务器里长出一个超级智能"没有。两者在能力落差上是同一件事,只是后者听起来像个软件更新。
格加维追问,为什么危险。马斯克的回答绕了一圈,核心是能力差距本身。一个比全人类加起来都聪明的存在,它的目标函数是不是和你一致,你没法验证,也没法在它变强之后再纠正。
外星人在哪
格加维顺势问费米悖论:你觉得地球之外有智慧生命吗?
马斯克说,看起来很有可能。他说这是物理学和哲学里最大的问题之一:外星人在哪。然后他开了个玩笑,说也许他们就在我们中间,有些人觉得我就是外星人,那不是真的。
格加维接了一句,在这个地区,人们相信人类本来就不来自地球,亚当和夏娃是从别处来的。所以某种意义上,人类对这片土地来说就是外星人。
马斯克没有展开这个宗教类比,转去做了一次粗略计算。任何一个对殖民银河系有兴趣的高级文明,即便不超过光速,只以光速的10%到20%航行,填满整个银河系也只需要一千万年,最多两千万年。
放在宇宙的尺度上,这不算什么。(马斯克)
这个计算的含义是:如果外星文明存在,他们早该到了。既然没看到,要么他们不存在,要么他们已经在这儿,而我们不够聪明,察觉不到。
地下五十层能解决堵车
格加维问起他发推特说要在华盛顿特区底下挖隧道。
马斯克先开玩笑,说这是秘密计划,别外传。然后正经讲了他的判断:城市拥堵的解法是城市地下的隧道网络。
他强调二维平面图不行,隧道得往深处走,做成多层立体网络。
你能往下挖的深度,永远超过你能往上盖的高度。最深的矿井比最高的楼还深。(马斯克)
所以理论上可以挖二十层、三十层、四十层、五十层,想挖多少挖多少。这样任何城市的拥堵都能解决。
真正的难点在工程:怎么把隧道挖得又快又便宜又安全。这三件事做到了,城市拥堵问题就消失了。他说这番话听起来很琐碎甚至有点傻,但他已经讲了很多年。
他对堵车的成因有个解释,很干脆。城市里的建筑是三维的,路网是一维的。而人们偏偏都想在同一时间进出这些建筑。三个条件叠在一起,必然堵。这个分析跟很多交通规划的思路不一样,后者喜欢讨论信号灯配时、潮汐车道、限行,他的解法是直接把路网从一维变成三维。
华盛顿、洛杉矶、美国几乎所有大城市,全世界几乎所有大城市,都卡在同一个地方。
颠覆会来,但需要二十年
格加维问,下一项会改变生活方式、思维方式、做生意方式的颠覆性技术是什么。
马斯克的答案是全自动驾驶汽车,而且会来得比大多数人以为的快。
但他马上给"快"做了限定。全世界大约有二十亿辆车,接近二十五亿辆汽车和卡车。全球新车年产能大约一亿辆。车的寿命是二十到二十五年。所以即使全自动驾驶明天就成熟,要替换掉现有车队也需要很长时间。
全自动驾驶出现的那一刻,并不是社会剧变的时刻。制造足够多的自动驾驶车辆去冲击就业,需要时间。这个冲击会在二十年里发生。(马斯克)
然后他补了一句更重的:二十年其实很短,如果这期间全球劳动力的12%到15%失业的话。
他还提了一句,驾驶可能是全世界雇佣人数最多的单一工种。这些人接下来做什么,需要新的规则。
给政府的三条建议
台下坐着139个国家的代表,格加维请他给政府官员三条建议。
第一条,盯紧人工智能。
马斯克的措辞很小心。他说需要非常谨慎地对待AI的采用,确保研究人员不会跑偏。他给的理由是科学家常见的一种状态:太投入自己的工作,反而不容易意识到自己做的事会带来什么后果。
从公共安全角度,政府需要密切关注人工智能,确保它不对公众构成危险。(马斯克)
第二条,认真考虑交通的电气化。
他说这个转变对很多原因都是好事,但会比自动驾驶慢,大致需要三十到四十年。
这里他给了一组能源数据:全球能源使用大致三分之一用于交通,三分之一用于供暖,剩下的是其他。随着这些转向电力,电力需求大概会变成现在的三倍。
所以真正要提前想的是:怎么发出这么多电。
第三条,准备应对大规模失业。
最终我们不得不有某种形式的全民基本收入。我认为没有别的选择。(马斯克)
他特意加了一句免责声明:这些不是他希望大家发生的事,是他认为大概率会发生的事。如果判断成立,那就得讨论怎么办。
钱的问题能解决,意义的问题不能
马斯克对自动化的前景其实挺乐观。产出会极高,几乎所有东西都会变得很便宜,那是一种丰裕。
所以全民基本收入在他看来是必然的,也是能办到的。难的是下一步。
更难的挑战是,人怎么获得意义。很多人从工作里得到意义。如果你的劳动不再被需要,那意义从哪来?你会不会觉得自己没用?这是难得多的问题。(马斯克)
他把这个问题抛在台上,没有给答案。整场对话里,这是他少有的明确承认"我不知道怎么办"的地方。
注意这里的逻辑顺序。大部分人讨论UBI是从分配公平切进去的,马斯克是从产出足够、但人的心理结构不匹配切进去的。他担心的不是钱不够,是人失去了被需要的感受。
我们已经是半赛博格了
对上面那个意义问题,马斯克给了一个很科幻的方向:生物智能和机器智能的融合。
他说人类其实已经部分是这个状态了。神经科学里有个三层结构:边缘系统,也就是动物脑、原始脑;皮层,负责思考和规划;再往上,是数字化的那个自己。
手机、电脑、各种应用,就是第三层。
他还提了个细节:一个人死了,他的数字幽灵还在。邮件、发过的照片、社交媒体,都还活着。
随时间推移,我们大概会看到生物智能和数字智能更紧密地融合。(马斯克)
瓶颈在带宽。计算机之间可以用每秒万亿比特通信,人的拇指大概每秒十比特,慷慨点算一百比特。输入方式的退化也很明显:以前用键盘,现在大部分输入靠两根拇指戳屏幕。
所以他判断,高带宽的脑机接口会是关键。它能实现人机共生,也许还能顺带解决AI的控制问题。
讲到这里他自嘲了一句:我讲得有点太玄了。
火箭复用的一笔账
格加维问他,为什么总是把人类的极限往前推。他拿火箭回收举例。
逻辑起点是飞机。假设飞机是一次性的,几乎没人坐得起飞机。一架747大概两三亿美元,往返要两架。没人会花几百万美元买一张机票。
但飞机能复用几万次,所以航空旅行变得便宜。
火箭同理。一枚火箭大约六千万美元。如果只用一次,这六千万就是单次成本。如果能用一千次,摊到每次是六万美元。再考虑到一次能载很多人,太空飞行的成本就能接近航空飞行的成本。
这件事难在地球重力。马斯克说,让火箭可复用的难度远高于让飞机可复用,因为地球重力井很深。他还补了一句:如果地球重力再强10%,这件事大概就不可能了。
能做成的前提是把材料、设计、每个细节都做到位。格加维问他失败了几次,他说七八次左右,然后用了一句典型马斯克式的修正:有四次是我愿意数的。
他招人只问两个问题
格加维问怎么选团队。马斯克说,说实话,主要是直觉。
面试时他问的问题基本固定。第一个是:
讲讲你的人生故事,讲讲你一路做出的决定,以及为什么那么决定。(马斯克)
第二个是:讲讲你解决过的最难的问题,你是怎么解决的。
他解释为什么这个问题有效。真正解决过问题的人,清楚自己是怎么解决的,知道那些小细节。假装解决过的人,只能讲一层,再往下就卡住了。
这个测试的好处是它不依赖面试官的领域知识。你不需要懂那个技术,只需要听对方能不能往深处走。一个真做过的人会主动提到自己踩过的坑、试错的路径、放弃的方案,这些细节编不出来。
他最大的挑战是维持一个纠错回路
格加维问他人生中最大的挑战是什么。他想了半天,说有很多,得先想想哪个最糟。
最后给了两个答案。一个是关于时间怎么花。另一个是:确保自己有一个能纠错的反馈回路,并且在很长时间里维持它,哪怕别人不愿意告诉你你想听的话。
他说这非常难。
这个回答跟后面的建议是同一条线。
别想着成为我
最后一个问题,给全世界的年轻人一句建议,特别是那些想成为马斯克的人。
马斯克先否定了问题本身:他们不该想成为我。
当我没有你想的那么好玩。(马斯克)
真正的建议是学物理的思维方式。不只是学公式,公式当然有用,更重要的是物理学家理解世界的那套框架,尤其是理解反直觉事物的方式。
第二条关于心态:始终假设自己在某种程度上是错的,目标是随着时间推移,错得越来越少。
他点出人类最常犯的错误是想当然。
你希望某件事是真的,哪怕它不是。于是你忽略掉真实情况,因为那和你希望的不一样。这个陷阱很难躲开,我自己也常掉进去。(马斯克)
第三条是主动去要批评,特别是跟朋友要。他的观察是,爱你的人希望你过得好,所以不想告诉你坏消息。你得主动问,并且明确表示你真的想知道,他们才会说。
这场对话里最值得带走的东西
回到开头那个火星的问题。马斯克的回答其实很克制:我不想去死,我只是不介意死在那里。这里面有一个思维习惯,值得单独拎出来。
他面对任何问题,先把它拆成几个可以量化的部分。火星死亡是一个情绪问题,他拆成"人总会死"和"选择一个地点"两件事,情绪就没了。堵车是一个感受问题,他拆成"建筑是三维的、路网是一维的、需求是同时的",解法就出来了。就业冲击是一个恐慌问题,他拆成"二十亿辆存量车、一亿辆年产能、二十五年寿命",就有了时间表。
这套方法他自己也说了,来源是物理。物理训练的不是公式,是一种反射:当结论和直觉冲突时,先怀疑直觉。
还有一件事值得留意。整场对话里,他对每一个预测都加了限定词。十年后大部分车是自动驾驶,但"快"指的是二十年。AI很危险,但他也说两个说法都对。五十年后的预测一定错,但他还是说了希望看到什么。这种表达方式在演讲场合很不讨巧,但它让那些判断经得起时间检验。
七年过去了,回头看这场对话,有几个判断已经应验,有几个还在路上,也有几个他自己后来改了主意。但那个提问的方式没变。
可以立刻用起来的几件事
- 判断一项技术会不会冲击就业,别只看技术成熟的年份。去查这个行业的存量资产数量和平均寿命。二十亿辆车加二十五年寿命,比任何专家的判断都可靠。
- 面试时问"讲讲你解决过的最难的问题,你是怎么解决的",然后听细节。讲不出三层细节的人,基本没真做过。这个测试不需要你懂那个领域。
- 把你最希望成真的那个判断写下来,然后专门去找反面证据。想当然的典型症状是所有证据都支持你,这时候最该警惕。
- 每个月主动找一个朋友问:"你觉得我哪里做得不对?"问完必须补一句"我真的想知道",否则对方会给你一个安全的答案。
- 学一套反直觉的思考框架。物理是最便宜的选择,重点不在公式,在"这个结论和我直觉相反,那我的直觉哪里出了问题"这个反射。
- 做长期预测时,先算量级再谈趋势。存量、产能、替换周期这三个数字算清楚,大部分热闹的预测会自动失效。
- 遇到一个让你焦虑的问题,试着把它拆成可以量化的部分。拆不动,说明你还没理解它。
0:00What a challenge! To talk to one of the most, in my opinion, in life, we've seen, within this part of the world, great people, like Al-Khawarizmi who invented algorithm. Globally, Newton. Henry Ford, the Wright Brothers, Albert Einstein and Elon Musk.
0:43See, you are in rush, You want to go to places that nobody has been. You are re-inventing a certain industry, from the rocket industry with SpaceX to the car industry with Tesla. What's your life mission? Why do you do whatever you do? Sure, first of all, thank you for having me.
1:12It's an honor to be here. And I'm having a really great time with my kids in Dubai. It's really been fantastic. I really encourage anyone who hasn't been to visit, what a great city! Thank you. And, in terms of the motivations, I used to like this sort of... kind of a long version of the explanation but...
1:42essentially, when I was a kid I was wondering, what's the meaning of life? Like, why are we here? What is it all about? And I came to the 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,
2:09the better we're able to ask these questions. And... So, I think there are certain things that are necessary to ensure that the future is good. And... some of those things are in the long term having long term sustainable transport and sustainable energy generation. And to be a space exploring civilization.
2:37And for humanity to be out there among the stars. And be a multi-planetary species. I think that being a multi-planetary species and being out there among the stars is important for the long term survival of humanity. And, that's one reason, kind of like life insurance for life collectively.
2:58Life as we know it. But then the part that I find personally most motivating is that it creates a sense of adventure, and it makes people excited about the future. If you consider two futures, one where we are forever confined to Earth until eventually something terrible happens.
3:21Or another future where we are out there on many planets, maybe even going beyond the solar system. I think that space invasion is incredibly exciting and inspiring. And there need to be reasons to get up in the morning. You know, life can't just be about solving problems. Otherwise, what's the point?
3:39There's got to be things that people find inspiring, and make life worth living. So, what is life for you? I mean, you look at our life, and I heard you before speaking. Is it a dream? Is it real? Is it a million deal? What is life for Elon Musk? I find that as I get older I find that question to be
4:10maybe more and more confusing or troubling or uncertain. Particularly when you see the advancement of something like video games. You know, 40 years ago, you had video games, the most advanced video game would be like Pong, when you had two rectangles and a dot. And you're like batting it back and forth.
4:34- I played it. - Yeah, me too, exactly. - Us all. - It sort of dates you a little bit. Yeah, we both played the same game. And that was like a pretty fun game at the time. But now, you can see a video game that's photo-realistic, almost photo-realistic, and millions of people playing simultaneously.
4:54And, and you see where things are going with virtual reality. And augmented reality and... if you extrapolate that out into the future with any rate of progress at all, like keeping 0.1% of something like that a year, then eventually those games will be indistinguishable from reality.
5:17They'll be so realistic you won't be able to tell the difference between that game and reality as we know it. And then, it seems like, well, how do we know that that didn't happen in the past? And that we're not in one of those games ourselves? Interesting. Interesting. I mean, it could be.
5:39Everything is possible in life. I mean there's... Yeah, particularly like things tend to be accelerating to something. Isn't it? I mean, if we look at our life, it seems in the past 100 years life has been accelerating quite fast. - Yeah. - In the past 20 years. - It's getting faster and faster.
5:58- Is it more slow? So, my question is really, how will life be 20, 30, 50 years from now? Our education, our transport. How do you see it? Well, I think this is one of those things that are quite difficult to predict. When you think of, say, the first controlled power flight was in 1903
6:23with the Wright Brothers. And then, 66 years later we put the first people on the moon. I mean, if you asked people, say, in 1900, what are the odds of landing on the moon they would've said that's ridiculous. If you try to talk to them about the internet they wouldn't know what the heck you're even...
6:48What are you talking about? Like, this sounds so crazy. But today, with a hundred-dollar device you can video-conference with anyone in the world. On the other side of the world, and if you have a Wi-Fi connection, it's basically free. You're free to have an instant visual communication with anyone,
7:13or even with millions of people. You know, with social media you can communicate to millions of people simultaneously. So, and you can google something and ask any question. It's like an oracle of wisdom, that you can ask almost any questions and get an instant response. It would be incredibly difficult to predict these things in the past.
7:33Even the relatively recent past. So, I think the one thing that we can be quite certain of is that any predictions we make today for what the future will be like in 50 years will be wrong That's for sure. I think directionally, I can tell you what I hope the future has, as opposed to maybe what it will be.
7:58This may just be wishful thinking. I mean I hope we are out there on Mars and maybe beyond Mars, Jupiter. I hope we're traveling frequently throughout the solar system, perhaps preparing for missions to nearby star systems. I think all of that is possible in 50 years. And I think it's going to be very exciting to do that.
8:31And, I think we'll see autonomy and artificial intelligence advance tremendously. Like that's actually quite near term. My guess is in probably 10 years, it will be very unusual for cars to be built that are not fully autonomous. - 10 years. - 10 years from now? Yeah. I think almost all cars built will be able of full autonomy
9:01in about 10 years. As it is, the Tesla cars that are made today, have the sensor system necessary for full autonomy. And we think probably enough compute power to be safer than a person. So, it's mostly just the question of developing the software and uploading the software. And if it turns out that more compute power is needed,
9:23we can easily upgrade the computer. And, so that's all Teslas built since October last year. And other manufacturers will follow and do the same thing. So, getting in a car will be like getting in an elevator. You just tell it where you want to go and it takes you there with extreme levels of safety.
9:46And that will be normal, that will just be normal. Like, for elevators, they used to be elevator operators. You get in, there will be a guy moving a lever. Now, you just get in, you press the button and that's taken for granted. So, autonomy will be wide-spread. I think one of the most troubling questions is artificial intelligence.
10:10And I don't mean narrow AI, like, vehicle autonomy I would put in the narrow AI class. It's narrowly trying to achieve a certain function. But deep artificial intelligence, or what is sometimes called artificial general intelligence, where you can have AI that is much smarter than the smartest human on Earth.
10:36This, I think, is a dangerous situation. Why is it dangerous? I mean, there are two views, one view is that artificial intelligence will help humanity, and there's another school of thought that artificial intelligence is a threat to humanity. - Why is that? - I think it's both.
10:58You know, it's like... one way to think of it is imagine we're going to be visited... imagine you're very confident that we're going to be visited by super intelligent aliens, in let's say 10 years or 20 years at the most. - Super intelligent. - So, you think within 20 years... - Yeah...
11:20- we'll have aliens on Earth? Well, digital super intelligence will be like an alien. - It will be like an alien. - Yeah. But my question is, do you think there is other intelligent life outside the Earth? It seems probable. I think this is one of the great questions in physics and philosophy,
11:45is, where are the aliens? Maybe they are among us, I don't know. Some people think I'm an alien. Not true. - Not true. - But maybe we are aliens. Maybe we aliens. I mean, if you look at this part of the world. Yeah. They believe that human beings are not from Earth, they came from somewhere else.
12:16Eve and Adam came from somewhere else to Earth. So, in a way, human beings are aliens to this land. Do you think we'll make contact with aliens within the next 50 years? Well, that's a really tough one to say. If there are super intelligent aliens out there, they're probably already observing us.
12:45That would seem quite likely and we're not smart enough to realize it. But I can do some back of the envelope calculations and... any advanced alien civilization that is at all interested in populating the galaxy, even without exceeding the speed of light, even if you're only moving at, say, 10 or 20 per cent of the speed of light,
13:18you could populate the entire galaxy in let's say 10 million years. Maybe 20 million years max. This is nothing in the grand scheme of things. Once you said you wanted to die on Mars. Why? To be clear, I don't want to die on Mars. It's like, if... we're all going to die someday,
13:44and if you're going to pick some place to die, then why not Mars? You know, if you're born on Earth, why not die on Mars? Seems like may be quite exciting. But, I think given the choice between dying on Earth and dying on Mars, I'd say, yeah, sure, I'll die on Mars. But it's not some kind of Mars death wish.
14:08And if I do die on Mars, I just don't want to go on impact. Let's come back to Earth, actually. You tweeted that you are building a tunnel under Washington D.C. Why? What is it? - It's a secret plot. - Okay. - Just between us. - Nobody helps you? Yeah, exactly, let's keep that a secret.
14:38I think this is going to sound a little... I mean, it seems like so much trivial or silly, but... I've been saying this for many years now but I think that the solution to urban congestion is a network of tunnels under cities. And when I say that I don't mean a 2-D plan of tunnels, I mean tunnels that go many levels deep.
15:07So, you can always go deeper than you can go up. Like, the deepest mines are deeper than the tallest buildings. So, you can have a network of tunnels that is 20, 30, 40, 50 levels, as many levels as you want, really. And so, given that, you can overcome the congestion situation in any city in the world.
15:35The challenge is how do you build tunnels quickly and at low cost and with high safety? So, if tunnel technology can be improved to the point where you can build tunnels fast, cheap and safe, then that would completely get rid of any traffic situations in the cities. And so, that's why I think it's an important technology.
16:04And, Washington D.C., L.A and most of the major American cities, most major cities in the world suffer from severe traffic issues. And it's mostly because you've got these buildings which are, these tall buildings that are 3-D and you have a road network that is one level. And then, people generally want to go in and out of these buildings at the exact same time.
16:27So, then, you get the traffic jam. Let's come back to... your year in Dubai. The first time I met you it was the 4th of June 2015, at your office in SpaceX. And, I asked you would you have a presence in UAE? And your answer was: I'm busy with China. Maybe not in the near future, and almost a year and a half later,
16:57we are here, seems time goes quite fast. Why now? I think actually things are going really well in China. So, we have some initial challenges figuring out charging and service infrastructure and various other things, but now it's actually going really well, and... so the timing seems to be good to really make a significant debut
17:25in this region, starting in Dubai. In your opinion, what is the new disturbing thing that will come next in technology? What's next in technology? - What's next in technology? - That will disturb the way we live, the way we think, the way we do business. Well, the most near to impact from a technological standpoint
17:54is autonomous cars, like fully self-driving cars. I'd say that's going to happen much faster than people realize. so, and that's... it's going to be a great convenience to be in an autonomous car, but there are many people whose job is to drive. So, if... in fact I think it might be the single largest employer
18:22of people is driving in various forms. And so, then we need to figure out new rules for what do these people do. But it will all be very disruptive and very quick. I should characterize what I mean by quick. Because there are... Quick means different things to different people. There are about two billion vehicles in the world.
18:52Approaching in fact 2.5 billion cars and trucks in the world. The total new vehicle production capacity is about a hundred million. Which makes sense, because the life of a car or truck before it's finally scraped is about 20-25 years. So, so the point at which we see full autonomy appear
19:18will not be the point at which there is massive societal upheaval, because it will take a long time to make enough autonomous vehicles to disrupt employment. So, that disruption I'm talking about will take place over about 20 years. Still, 20 years is a short period of time to have I think something like
19:4412 to 15 per cent of the world force be unemployed. Thank you. This is the largest global government summit we have over 139 governments here. If you want to advise government officials to be ready for the future, what three pieces of advice can you give them? Well, I think the first bit of advice is to really play close attention to...
20:17the development of artificial intelligence. I think this is, we need to be 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 that scientists can get so engrossed in their work, they don't necessarily realize the ramifications of what they're doing.
20:44So, I think it's important for public safety that we... you know, governments keep a close eye on artificial intelligence and make sure that it does not represent a danger to the public. Let's see, secondly I would say we do need to think about transport in general. And, there's the movement towards electric vehicles,
21:18sustainable transport, I think that's going to be good for many reasons, but again, not something that happens immediately, that's going to happen slower than the self-driving vehicles. Because that's probably something that happens over 30 or 40 years. The transition to electric vehicles.
21:41So, thinking about that in context... the demand for electricity will increase dramatically. So, currently, in terms of total energy usage in the world, it's about 1/3, about 1/3 transport, about 1/3 heating. So, over time that will transition to almost all... not all, but predominantly electricity, which means that the demand for electricity will probably triple.
22:17So, it's going to be very important to think about how do you make so much more electricity And... It seems they'll have an easy job, that's it, there are no more challenges for them. No, well, I think maybe... these things do play into each other a little bit, but what to do about mass unemployment?
22:44This is going to be a massive social challenge. And I think ultimately will have to have some kind of universal basic income I don't think we're going to have a choice. - Universal basic income. I think it's going to be necessary. So, it means that unemployed people will be paid across the globe.
23:05- Yeah. - Because there are no jobs. Machines, robots are taking over. There will be fewer and fewer jobs that a robot cannot do better. That's simply... And I want to be clear, these are not things that I wish would happen. These are simply things that I think probably will happen.
23:27And so, if my assessment is correct and they probably will happen, then we need to say what are we going to do about it. And I think some kind of universal basic income is going to be necessary. Now, the output of goods and services will be extremely high. So, with automation, there will come abundance.
23:52There will be... almost everything will get very cheap. The... So... I think the biggest... I think we'll just end up doing a universal basic income. It's going to be necessary. The harder challenge, much harder challenge, is how do people then have meaning? Like a lot of people they derive their meaning from their employment,
24:22so, if you don't have... if you're not needed, if there's not a need for your labor, how do you... what's the meaning? Do you have meaning? Do you feel useless? That's a much harder problem to deal with. And then how do we ensure that the future is going to be the future that we want?
24:43That we so like. You know, I mean do think that there's a potential path here which is, I'm really getting into science fiction or sort of advanced science stuff. But, having some sort of merger with biological intelligence, and machine intelligence. To some degree, we are already a cyborg.
25:20You think of like the digital tools that you have, your phone and your computer, the applications that you have. Like the fact that as I mentioned earlier you can ask a question and instantly get an answer from Google or from other things. And, and so you already have a digital touchery layer.
25:41I say touchery because you can think of the limbic system, kind of the animal brain or the primal brain and then the cortex, kind of the thinking, planning part of the brain, and then your digital self as a third layer. So, you already have that, and I think if somebody dies, their digital ghost is still around.
26:01All of their e-mails and the pictures that they posted and their social media. That still lives, even if they died. So, over time I think we'll probably see a closer merger of biological intelligence and digital intelligence. And it's mostly about the bandwidth, the speed of the connection between your brain and your digital...
26:28the digital extension of yourself. Particularly output, and, if anything is getting worse, you know, we used to have keyboards that we used a lot, now we do most of our input through our thumbs, on a phone. And, that's just very slow. A computer can communicate at a trillion bits per second, but your thumb can maybe do maybe 10 bits per second or a hundred
26:57if you're being generous. So, some high bandwidth interface to the brain I think will be something that helps achieve symbiosis, between human and machine intelligence and maybe solves the control problem and the usefulness problem. I'm getting pretty esoteric here, I don't know is this is...
27:25It's close, we got it. Always you think out of the box. Your ideas are so huge. You want to go to space, you decided to go to space, you did it. You decided that you wanted to land your rocket back, - you failed, 7 times, 8 times? - Yeah, something like that. - Then it landed. - 4 times that I care to count.
27:55How do you come with these ideas? Sometimes they are pushing the human limit. You are always pushing the human limit, why? Well, I... I think about what technology solution is necessary in order to achieve that particular goal, and then try to make as much progress in that direction as possible.
28:20So, in the case of space flight, the critical breakthrough that's necessary in space flight, is rapid incomplete reusability of rockets. Just as we have for air crafts. You can imagine that if an air craft was a single use, almost no one would fly. Because you can buy like, say, 747 might be...
28:46250 million Dollars, 300 million Dollars, something like that. You need two of them for a round trip. But nobody is going to pay millions of Dollars for a ticket to fly. To do air travel. So, but because you can re-use the air craft tens of thousands of times, the... Air travel becomes much more affordable.
29:11And, the same is true of rockets. Our rocket costs... 60 million Dollars, roughly. So, a capital cost if it can be used once in 60 million Dollars. But if the capital cost if it can be used a thousand times is 60 thousand Dollars. So, then if you can carry a lot of people for a flight,
29:38then you can get the cost of space flight to be something not far from the cost of air flight. So, it's truly fundamental, because earth gravity is quite deep. Earth has a fairly high gravity. The difficulty of making a rocket reusable is much greater than the difficulty of making an air craft reusable.
30:02That's why a fully reusable rocket hasn't been developed that far. But if you use the most advanced materials, the most advanced design techniques, and you get everything just right, then I'm confident that you can do a fully reusable rocket. Fortunately, if Earth gravity was even 10 per cent stronger,
30:25I would say it wouldn't be impossible. You need a team around you to deliver a lot of ideas. How do you choose your team? Based on what? Well, I suppose honestly that it tends to be a gut feeling more than anything else. So, when I interview somebody, the main questions are always the same.
30:54What do you ask? I say: Tell me the story of your life. And, the decisions that you made along the way and why you made them. And then, and also tell me about some of the most difficult problems you worked on and how you solved them. And, that question I think is very important, because...
31:18the people that really solved the problem, they know exactly how they solved it. They know the little details. And the people that pretended to solve the problem, they can maybe go one level and then they get stuck. So, what was your biggest challenge in life? Biggest challenge in life?
31:50- No challenge? - Well, no, there's a lot of them. I'm trying to sort which is the worst. I think just thinking about how to spend time. One of the biggest challenges I think is making sure you have a corrective feedback loop, and then maintaining that corrective feedback loop over time, even when people won't to tell you exactly what you want to hear.
32:23- That's very difficult. - Yes. Time is over. I'll ask you just one last question. If you allow me. In the World Government Summit we have so many people from... so many young people actually from across the Globe. If you have an advice to them, young people globally who want to be like Elon Musk.
32:47What's your advice to them? I think that probably they shouldn't want to be. - You? - I think it sounds better than it is. Okay. Yeah, it's not as much fun being me as you'd think. - I don't know. - You don't think so? It could be worse, for sure. But it's... I'm not sure I want to be me.
33:18Okay. But... You know, I think my advice is if you want to make progress in things, I think that the best analytical framework for, I'll say in the future is physics. I'd recommend studying the thinking process around physics. Like, not just the equations, the equations are certainly very helpful,
33:52but the way of thinking in physics, it's the best framework for understanding things that are counter–intuitive. And, you know, always take the position that you are to some degree wrong, and your goal is to be less wrong over time. One of the biggest mistakes people generally make and I'm guilty of it too is wishful thinking.
34:19You know, like you want something to be true, even if it isn't true. And so you ignore the things that... You ignore the real truth, because of what you want to be true. This is a very difficult trap to avoid. And like I said, it's certainly one that I find myself in, having problems with.
34:42But, if you just take that approach of that you're always to some degree wrong and your goal is to be less wrong. And solicit critical feedback, particularly from friends. Like, friends, if somebody loves you they want the best for you. They don't want to tell you the bad things.
35:01So, you have to ask them and say: I do really want to know. And then they will tell you. Thank you very much. It's been... It's great for the World Government Summit to have a legend, who's creating the future for humanity, to share his thoughts, his ideas, his visions, challenges,
35:35and his hope for life. Thank you very much. Thanks for having me.
资源与出处
- 视频在 YouTube 收看
- 档案silicon-valley.tech 档案页本页视频与文稿数据来自 Elon Musk Archive
- 精读本篇精读 Markdown 版