这场对话的真名,叫“如何拯救人类”
台上摆着两杯水。马斯克指了指:这杯是热水,这杯是常温水,你随便挑。
然后迈克尔·米尔肯(Michael Milken)拿起话筒,念了一段议程上没写的细节。他说,官方标题写的是“与埃隆·马斯克的对话”,可这场对话真正的名字叫“如何拯救人类,以及其他类似的话题”。
接下来一个多小时,马斯克从外星人讲到出生率,从加州高铁讲到 AI 会不会撒谎。几乎每个话题最后都落在同一个坐标上:人类文明能撑多久,靠什么撑。
这场对话信息密度相当高。下面把它拆开讲。
11 年前列的五件事,现在应验了几件
米尔肯先放了一段录像。2014 年,同一个舞台,画面里的马斯克还年轻一些。
当时他被问到:大学时代你认为什么会最深刻地影响人类未来。
他列了五件事。互联网;可持续能源,太阳能负责生产端,电动车负责消费端;人类成为多行星物种;修改人类基因组;人工智能。
关于基因组那条,他当场补了一句澄清:他只是说这件事“会产生巨大影响”,并不代表“应该去做”。预测和主张,他分得很清楚。
今天回头看,五件事里三件已经从科幻变成了产业。互联网重塑了信息分发,特斯拉把电动车推成了主流,可重复使用火箭让多行星第一次有了工程路径。基因编辑有了 CRISPR。AI 从论文模型变成了能写代码、能对话、能生成视频的东西。
“科幻早就在想这些了。我想做的是,让科幻不要永远是虚构的。”(马斯克)
这句话听着像口号。但他接着解释了这个执念的来源。
他说,人活着得有让心跳加速的东西。如果每天醒来面对的都是“解决一个又一个问题”,生活会变得很沉重。太空探索这类事情的意义,在于给未来一个值得期待的方向。
他还讲了一个观察:随便问世界上任何一个五六岁的孩子,什么最让你兴奋,答案里几乎一定有太空。这是跨文化的。
所以他不希望阿波罗计划成为人类探索的最高水位线。
他提到自己当年写过一封信,毛遂自荐想去运营阿波罗计划。米尔肯笑着说“你会干得很棒”,马斯克自己也笑了。
信是次要的。他的判断是:一旦探索的天花板停在上世纪六十年代,人类就会在心理上认定“我们最好的时候已经过去了”。他不接受这个前提。
费米悖论:外星人在哪,为什么一颗都没碰到
聊到太空,绕不开一个问题:外星人到底在哪。
马斯克先把时间尺度摆出来。宇宙大约 138 亿岁,地球大约 45 亿岁。一个文明哪怕能持续一百万年,放在 138 亿年里也只是小数点后面很靠后的数字,几乎可以忽略。
人类文明怎么算?他从最早的文字开始算。苏美尔人的楔形文字大约出现在 5500 年前。5500 年对比地球 45 亿年的寿命,是百万分之一。
换句话说,人类有文字的历史,只占地球寿命的百万分之一。
“如果一个文明能持续一百万年,那已经非常了不起了。而如果能持续一百万年,它大概率会散布到银河系的每个角落。”(马斯克)
这就是费米悖论(Fermi Paradox)。意大利物理学家恩里科·费米(Enrico Fermi)当年问的那句“他们都在哪儿呢”,到今天依然没有答案。
马斯克提到一个现象:很多人相信外星人就在我们中间。
“有意思的是,这批人里很多同时也相信人类从来没登上过月球。你想想这个逻辑。”(马斯克)全场笑了。
他自己怎么判断?他说他没见过任何外星人存在的证据。SpaceX 的星链(Starlink)星座大约有 6000 颗卫星,从来没有一次因为要避让 UFO 而机动。
“从来没有出现过‘嘿,那是什么,是外星人吗’这种情况。”(马斯克)他的结论是:如果谁有外星人存在的证据,只要不是模糊的一团光斑,他很愿意看看。但他倾向于认为没有。
这就引出了他对文明的核心判断。
银河系里任何一个文明只要持续一百万年,哪怕以远低于光速的速度前进,比如光速的百分之几,也足够探索并殖民整个银河系。既然没有,那说明什么?
“答案可能是,也可能大概率是:文明是脆弱而稀少的。你应该把人类文明想成广袤黑暗里的一支小蜡烛,我们要尽一切可能,别让这支蜡烛熄灭。”(马斯克)这句话是整场对话的情感锚点。后面聊监管、聊出生率、聊 AI,逻辑都从这儿长出来。
多行星的意义,是给文明买一份行星级保险
很多人误解马斯克的火星计划,以为他觉得地球没救了,要跑路。他专门澄清了这一点。
“重点不是从地球搬到另一个星球,然后让地球死掉。我完全不是这个意思。我要的是行星冗余,让任何单一事件都没法终结人类文明。”(马斯克)这就像一个公司不会把所有数据放在一台服务器上。地球现在就是那台唯一的服务器。
他提到费米悖论里的一类解释,叫“大过滤器”(Great Filter)。一个文明要持续存在,得跨过若干道关卡。不成为多行星文明,等于在原地等着某一天被自伤或者自然灾害清场。
恐龙就是被一颗大陨石清掉的。
再往远看,太阳终有一天会膨胀到吞掉地球,人类会被烤化。这件事一定会发生,只是时间问题。中间还有更近的风险。所以他的排序是:先跨过“单行星文明”这道过滤器,其他讨论才有空间。
要实现这件事,底层发明只有一个。
“人类成为多行星物种所必需的根本性发明,是快速可重复使用、可靠的火箭。”(马斯克)他故意模仿海盗的腔调念了这句,然后补了一句“太空海盗,为了胜利”。
聊到“你想死在火星上吗”这个老问题,他的回答很马斯克。
“想过。但后来我考虑了撞击身亡这个边界情况。”(马斯克)紧接着他说,真愿意死于撞击的话,今年就能实现这个目标。
全场笑。但这句话背后的意思很清楚:技术路径早就通了,卡住的是安全性和可靠性,想象力早就够了。
言论自由:宪法保护人民,约束的是政府
米尔肯拿出一叠马斯克过去的发言,让他逐条评论。第一条关于言论自由。
“言论自由是民主的基石。没有它,美国就完了。”(马斯克)他的论证很直接:如果人们拿不到判断候选人和政党所需的信息,就不可能做出知情的民主选举。言论被根本性地限制,就别指望选民做出正确决定。
他进一步问:为什么言论自由是第一条修正案?
因为当年的移民来自那些“说错话就会被关起来或者杀掉”的国家。他们吃过这个亏,所以把这一条放在最前面。
“世界上很多地方,你说不出自己真正想说的话,除非愿意承担后果。”(马斯克)他提醒在场的人,宪法的作用是保护人民免受政府侵害,它的设计初衷就是让“改变”变得困难。
“人们有时候会忘记宪法为什么在那儿。它是用来防政府的。”(马斯克)米尔肯接了一句“别把它当成理所当然”。这句话放在今天,分量不轻。
社会主义的根本错误:把资本配置权交给了效率最低的手
第二条引言更尖。
“社会主义的根本错误,是把资本配置权从高效的企业家手里,转移到效率惊人的政府手里。”(马斯克)他说这句话在这个房间里应该能找到不少共鸣,全场鼓掌。
他的论证方式有点意思。常见的说法是“不能让贪婪的公司来干,应该让政府来干”。他的回应是:政府本质上就是一家公司,而且是终极形态的公司。
“政府就是拥有暴力垄断权的公司。如果你不放心商业公司来做,那你更应该不放心政府来做,因为政府是最企业化的那种企业。”(马斯克)他补了一个生活化的类比:你在公司里还能通过买股票、换工作、投诉来影响结果,在政府面前几乎没有这种杠杆。
“所有人都去过车管所(DMV)。你想要一个放大版的 DMV 吗?大概不想。”(马斯克)这句话后来被米尔肯重复了一遍,成了全场的一个梗:“政府就是放大版的 DMV。”
除了 merit,别的标准一旦开口就收不住
第三条引言关于择优。
“基于 merit 之外的任何东西搞歧视,都是错的。”(马斯克)他的逻辑是滑坡:一旦你开始按非能力标准做筛选,就不知道在哪里停。
“我们得尽可能严格地按能力来,这在我看来是基础性的东西。”(马斯克)他没有展开太多,但这条价值观在后面反复出现:一个系统如果奖励的东西跟“做得更好”脱钩,它迟早会失去纠错能力。
米尔肯打趣说“很高兴你同意你自己”,马斯克回了一句“我也在想这人是谁,说得挺对”。
法规不会死,人死了法规还在
第四条引言是整场对话里最有洞察力的一段。
“监管和监管的一致性,就像格列佛被几千根小绳子绑住。我们是一条规定一规定地失去自由的。”(马斯克)他接着讲了一个很多人没意识到的点:法律和监管是永生的。人会死,法规不会。
“如果年复一年地通过更多法律、更多监管、设立更多监管机构,最终一切都会变成非法。”(马斯克)他举的例子是加州高铁。花了几十亿美元,修出来一小段,连铁轨都还没铺上。
“加州现在几乎把一切都变成非法了。所以你没法推进任何事情。”(马斯克)那历史上是怎么清理这些监管蛛网的?
他的回答很直接:战争。战争会强制性地推倒旧规则。
“我们当然不希望有战争。但要让文明在没有战争的情况下继续运转,你就必须主动淘汰法律和监管。你需要一个针对规则的垃圾回收机制。”(马斯克)没有这个机制,文明就会动脉硬化,一年比一年难做事。
米尔肯补了一句:过去还有日落条款(sunset clause),法规到期自动失效。现在这种做法很少见了。
今天上课的方式,像一百年前的杂耍
聊到教育,马斯克的比喻很传神。
“今天教育运作的方式,很像 vaudeville(杂耍剧场)。”(马斯克)他解释:在电影、电视、游戏出现之前,每个小镇都有自己的剧团,那就是当地的娱乐。大城市演员好一点,小城市差一点,但大家都只能看本地的。
后来电影来了。你把全国最会演戏、最会写剧本、最会做特效的人凑到一起,花几千万甚至上亿美元,做一部电影或者一个游戏。
结果就是:如果你不在纽约,而在贝克斯菲尔德,你看到的蝙蝠侠是当地剧团演的版本,不会是好莱坞那个版本。
“今天上课的方式就是这样。”(马斯克)一个老师站在黑板前讲课,本质上就是地方剧团的水平,和全世界最顶级的互动内容没法比。
他的主张是:把学习做成尽可能有吸引力的互动体验,有实时反馈,而不是让老师对着黑板演杂耍。
但他强调,游戏化只是手段,教育还有几条更根本的原则。
第一条,建立相关性。
“人脑在进化上把记忆看得很贵,它一直在尽量忘掉东西。你只有让大脑建立相关性,或者带强烈情绪,它才会留下来。”(马斯克)所以教一门课,得先告诉孩子它为什么重要。否则学完就忘。
第二条,教问题,别教工具。
他举了一个修汽车发动机的例子。你说“我们来搞清楚这台发动机怎么工作,把它拆开”。要拆,你需要扳手、螺丝刀、吊机、滑轮。拆的过程中,你自然学会了这些工具。
“这是有吸引力的。但通常的教学方式是什么?给你开一门螺丝刀课,再开一门扳手课。你会问,我为什么要上一门扳手课?”(马斯克)这就像教微积分却不告诉你微积分能用来干什么。学完,忘了。
合法移民慢得让人崩溃,非法移民快得离谱
米尔肯说接下来聊一个“没有争议的话题”,移民。全场笑。
马斯克先亮立场:他非常支持扩大合法移民,欢迎有才华、肯努力、诚实的人。他反对的是大规模、不经审查的移民。
“荒谬的是,合法移民美国又难又慢,非法入境却又容易又快。这显然说不通。”(马斯克)他说自己经历过合法移民流程,9·11 之后越来越糟,整个过程像卡夫卡小说,漫长又荒诞。他有朋友连给妻子办绿卡都办不下来。
另一边,从南部边境过河极其容易。
他说自己亲自去过边境,想看看是真是假,是不是宣传。到了之后发现是真的。
“人像《僵尸世界大战》(World War Z)一样涌过边境。我当时想,我们到底在检查谁?这不太健康。”(马斯克)他的结论分两半:大幅加速合法移民,同时保证南部边境有基本审查。这两件事并不冲突。
“我是移民的坚定支持者。但大规模不审查的移民,是灾难的配方。”(马斯克)星链真正的价值:把几十亿人接进能自学的世界
聊到星链(Starlink),马斯克把它和教育直接连了起来。
他的判断是:只要能上网,世界上的信息几乎都能免费拿到。麻省理工(MIT)把所有课程都放在线上了。问题在于你得先有网。
“有网络接入,你就能拿到全世界的信息。没有网络,或者太贵、带宽太低,你就拿不到 MIT 的课,也没法把你生产的东西卖出去。”(马斯克)他把这件事提到减贫的高度:互联网连接可能是把人们拉出贫困最有效的技术之一。有了连接,人们可以自学任何东西,也可以把商品和服务卖到全世界。
他还用了一个经济学框架来解释为什么这会影响一个国家的 GDP。
“GDP 是人均生产力的函数。如果有技术能提高人均生产力,你会在 GDP 上看到它。”(马斯克)星链在他看来就是这种级别的技术,能让一些国家的 GDP 直接抬一个台阶。
文明会终结,我们只是不希望它来得太早
第五条引言关于文明脆弱性。
“我们应该始终把文明视为脆弱的。不存在一条必然向上的轨迹。”(马斯克)他建议在场的人都去读历史。这句话听起来像常识,但他给了一个具体的例子。
苏美尔人的语言被遗忘了很久。那些刻着字的泥板,几千年里没人看得懂,直到 19 世纪才被破译。
“那就是一个早已死去的文明的遗迹。而这样的文明,有很多。”(马斯克)他的态度谈不上悲观,更像一种清醒。文明会结束,这是确定的。不确定的是时间。
“我们的文明终有一天会结束。我们只是不希望它来得太早。”(马斯克)AI 是全场最重要的问题,人类智能的占比会跌到 1% 以下
观众提问环节,第一个问题就是 AI 会怎么影响日常生活。
马斯克停了一下,说这可能是所有问题里最重要的一个。
“生物智能占全部智能的比例,每个月都在变小。最终,生物智能的占比会低于 1%。”(马斯克)这话听着吓人。他马上补了一个转折,说自己真正想讲的是另一件事:不要造一个脆弱的 AI。
他给了一个很工程的理由。硅基电路板在自然环境下表现不好,风吹日晒、温度变化、辐射,都会出问题。生物智能可以当备份,当缓冲。
“但作为比例,几乎所有智能都会是数字的。那我们的角色是什么?我不知道。”(马斯克)这是他少有的公开承认“不知道”。
他真正在意的是 AI 该怎么造。他想了很久 AI 安全,给出两条原则。
第一条,把 AI 做成最大程度追求真相的。
“AI 不应该被教着撒谎。哪怕真相在政治上不正确,它也应该说出它认为真实的东西。”(马斯克)他拿《2001 太空漫游》举例。HAL 9000 杀宇航员,动机来自它被要求同时保守一个秘密、又必须把宇航员送到目的地。它被迫撒谎,逻辑崩了。
“阿瑟·克拉克想说的就是:别让 AI 撒谎。很多人没意识到这一点。”(马斯克)第二条,把 AI 做成最大程度好奇的。
他的推理是:一个追求真相、充满好奇的 AI,更可能去保护和培育人类文明,因为人类比一堆石头有意思得多。
“火星的渲染很简单,看起来就像亚利桑那沙漠的一块。但人类文明的渲染复杂度要高好几个数量级。一个好奇的 AI 会更愿意看看人类文明能走到哪一步。”(马斯克)SpaceX 几乎不用 AI,因为 AI 在这些问题上还很糟
有人问 AI 能不能加速他的太空计划。
马斯克的回答有点出人意料。
“说来奇怪,太空探索几乎是 AI 用得最少的领域。SpaceX 基本不用 AI,星链也不用。我不反对用它,只是还没找到用处。”(马斯克)他说自己会拿 AI 问费米悖论、问火箭发动机设计、问电化学。
“到目前为止,AI 在这些问题上表现都很糟糕。还有很长的路要走。”(马斯克)这段话和前面“生物智能占比会低于 1%”放在一起看,意思就清楚了:AI 的长期影响巨大,但当下它在硬工程问题上的能力被高估了。
出生率跌下去,文明可能在成人纸尿裤里咽气
最后一个问题关于出生率。马斯克说这件事困扰他很久。
他给了一个估算人口的简单方法:看一个国家去年出生了多少婴儿,乘以预期寿命,就是这个国家未来的稳态人口。
“出生率并不是恒定的,它在下降。而且看二阶导数,下降还在加速。出生率的二阶导数非常糟糕。”(马斯克)他判断的后果很重。
“这不会通向一个更伟大的文明。它通向一个可能不会轰然倒塌,而是在成人纸尿裤里呜咽死去的文明。那是个悲伤的结局。”(马斯克)他举了韩国的例子。韩国生育率曾经是 6,现在跌到 0.75 左右。
他建议在场的人至少生三个孩子。他的原话是“孩子总得从某个地方来”。
然后他讲了一段很少被提到的历史。
“大约公元前 50 年,罗马通过一项法案,给任何生第三个孩子的罗马公民发奖金。也就是说,公元前 50 年罗马的出生率就已经是个问题了。罗马人不制造罗马人了。”(马斯克)古希腊也一样。公元前 800 年到公元前 300 年左右,希腊人生育率高,存活率高,所以希腊城邦遍布地中海。后来也掉下去了。
他给出的解释有点反直觉:繁荣会摧毁出生率。
“当一个文明感觉不到有意义的外部威胁,又非常繁荣,出生率就会暴跌。你会以为资源多了应该生更多,事实正好相反。”(马斯克)米尔肯补充了学界常提的三个因素:繁荣、医疗进步、女性教育。1900 年,全球一半的孩子活不过五岁。
这两组解释放在一起,指向同一个难题:让出生率回升,不能靠退回贫困,只能靠重新设计激励和文化。
让他失眠的,和孩子给他的快乐
最后一个问题是“什么让你夜不能寐,什么给你快乐”。
“孩子给我最多的快乐。我并不是说这是生孩子的理由,孩子本来就该生。但孩子确实是我人生中最大的快乐来源。”(马斯克)至于失眠,他说是任何文明级别的风险。出生率持续下跌、美国民主的基础被削弱、社会偏离择优系统,这些都会让他睡不着。
然后他讲了一个细节。
“我其实是听着文明衰落的播客入睡的。可能这就是问题的一部分。”(马斯克)他推荐了两个播客,《文明的衰落》(Fall of Civilizations)和《硬核历史》(Hardcore History)。他说自己基本是听着历史播客睡觉,所以晚上才会反复琢磨这些事。
把这场对话拆成能用的东西
这场对话表面上跨度很大,从外星人到出生率。底层其实是一条线:文明是脆弱的,能延续多久取决于我们主动做了什么。
马斯克给出的东西里,情绪成分很少,判断框架成分很多。把这个框架落到日常,有几件事是普通从业者今天就能做的。
- 用“蜡烛模型”审视你所在的系统。你的团队、公司、产品,有没有单点故障?有没有行星级冗余?把关键依赖列出来,至少给最重要的那一两个准备备份方案。
- 遇到“该让政府或者大公司来做”的争论,先问一句:这个主体的激励机制是什么?马斯克的 DMV 类比可以当成一个思维工具,用来判断任何组织。
- 给自己负责的流程做一次“监管垃圾回收”。找出那些没人记得为什么存在、但一直在拖慢速度的规则,列出来,主动申请废止或者设置日落条款。
- 设计培训或教学内容时,先讲问题再讲工具。别一上来就教螺丝刀,先给一台要拆的发动机。学完能用的知识,才是能留下来的知识。
- 用“出生数乘预期寿命”这个公式,快速判断任何一个国家或行业未来二十年的人才供给。做长期规划时,这个数字比当下的热度可靠得多。
- 如果你在用 AI,优先检查它有没有被鼓励说真话。一个被训练成迎合的模型,长期看比一个能力弱却诚实的模型更危险。
- 给自己留一点“仰望”的时间。马斯克反复强调,人需要让心跳加速的目标。目标不一定是火星,但得有一件事让你觉得未来值得期待。
这场对话最值得记住的画面,可能就是马斯克反复回到的那一支小蜡烛。
它没有说人类一定会赢。它说的是,这件事需要有人主动去护着。
0:00Well, these lights are bright. [laughter]
0:04See, you know, a lot of times you say things, me or you, and they edit them. So, I noticed this panel description was a conversation with Elon Musk. The real title of this panel was how to save the human race and other like topics. a conversation with Elon Musk. And so before we begin, Elon, I thought we might want to go back in time 11 years when you were sitting on this stage.
0:41That's hot water and this is regular. You can take your choice.
0:44All right, sounds good. Okay. So 11 years ago, Alain was talking about the things that will have the biggest impact on the future of humanity that he was thinking about in college. And so let's run that video of 11 almost 11 years ago to today. I guess you know when I was in college that I I thought about things that would most affect the future of humanity and and there were three areas that I thought would have the biggest impact and those were the internet, sustainable energy of which solar power is the production side and electric cars the consumption side and then
1:33humanity becoming a multifunet species. And so we cut that short, but there was two others you talked about. One was modifying the human genome.
1:44Yeah. I'm just I'm not saying we should. I'm just saying that that's the thing would really affect the future.
1:48And next AI. So a lot of people didn't weren't thinking about these same five things when they were in school. particularly humanity on multilanets at that time. Well, sci-fi was certainly thinking about it, but you know, I think at some point we want to make science fiction not fiction forever. And yeah, so like let's make life multilanetary and be a space fairing civilization, be out there among the stars. you know, I think there are things that like you have to be excited about the future.
2:23life life cannot just be about solving one problem after another. they have to be things that that that really sort of move your heart and that make you excited to wake up in the morning. and I think being becoming a space bearing civilization is one of those things. If you ask kids anywhere around the world, like what is what are some of the most inspiring things you can ask a like 5-year-old, six-year-old anywhere in the world and they're going to say, you know, space exploration is one of those things. and and we want to make sure that we we're we're you know that Apollo is not the high watermark. in fact, you mentioned at one point that that you wrote a letter offering to
3:02run the Apollo program, I believe.
3:05Yeah. [laughter] Long but but and I would have you would have done a fantastic job. But the the point is that the Apollo program was something that was inspiring to everyone around the world and we we don't want the Apollo program to be the high watermark of of human exploration. And yeah, we want I think you want to have some some sense that the future is going to be better than the past. That we're going to be out there going to other star systems and you know what what you see in a science science fiction non-dystopian sci-fi story of which there are not many but like Star Trek I suppose. Well, speaking of Star Trek, Alan Elon, when I think
3:49about you, let's look at Spocks from Star Trek here. Space, the final frontier. These are the voyages of the starship Enterprise. her ongoing mission to explore strange new worlds. To seek out new life forms and new civilizations, to boldly go where no one has gone before.
4:24So when I think about you, I think about Spock and Captain Kirk, and you're going to take us to places we've never gone before.
4:32Yeah, that's the idea. [applause] you know, if we if we send probes out there, we might I mean, we we might find the remains of long dead alien civilizations. if physics is correct, the the universe is about 13.8 billion years old. Earth is about 4.5 billion years old. So, but at 13.8 billion years, a civilization that even lasted a million years is three digits past the decimal point. And if you consider human civilization, I I I date it from like the first writing. So that first writing was the ancient Samrians archaic preuniform around 5,500 years ago. So that is 1
5:17millionth of Earth's lifespan. That's how long writing has existed. So if we were to last if if as a civilization for a million years, that would be incredible. and we would actually probably be in every part of the galaxy. So this is this causes me to to think that well where are the aliens? It's the Fermy question. You know the the great physicist Italian physicist Enrico Fermy he he's like where are they? now a lot of people think there are aliens among us.
5:52well there was that there was that movie Men in Black. Yes. You know they
5:57Yes. Yes. told is there among us and Elvis really went back to his own planet.
6:03Yeah. well I mean really a lot of people think there are aliens but I get asked that a lot and for some reason this the a lot of the same people who think there are aliens among us didn't think we don't think we went to the moon. which I'm like think about that for a second. [laughter] you know, so but I I I I think I would I mean if I've not seen any evidence of aliens and spa SpaceX with
6:31the Starling constellation has roughly 6,000 satellites and and not once have we had to maneuver around a UFO. Okay. [laughter] So So we were like, "Hey, what's that? Is that an alien?" Has occurred never. [laughter] so, so I'm like, okay, I don't see any evidence of aliens. And, I look at it, I if somebody has evidence of aliens in a in a, you know, that's not just a fuzzy blob, then I'd love to see it.
7:04Love to hear about it. and, but I don't think there is. So that which is actually reasonable concern because you could if if any civilization in the Milky Way in our galaxy were to last for a million years even with a speed of travel that's far below the speed of light you know like a few% of speed of light they could easily have explored and colonized the whole galaxy. So so they haven't so why not? I think the
7:37the the answer might be or perhaps probably is that that civilization is precarious and rare. and that we you should really think of human civilization as being like a tiny candle in a vast darkness and we should do everything possible to ensure that that candle does not go out.
8:00[applause] So Alan, I thought one of the interesting things for the people on X viewing this session and the people in the audience here [laughter]
8:14is that maybe I'd give you a few of your quotes and you can comment on them.
8:19Okay, let's start with this one.
8:21Free speech. Freedom of speech is the bedrock of democracy. Without it, America ends. Yes, it's it's obviously not possible to have democratic elections if people do not have access to the information that would allow them to make the right decision on a candidate or a party. So if if speech is constrained in a fundamental way, you you just can't expect people to make the the right decision or an informed decision because they are prevented from being informed.
8:56I think it's it's it's very a foundational element. It's you know I have to say like why why is free speech freedom of speech the first amendments because people came from countries where if you spoke freely you would be imprisoned or killed. That was why they were like you know what we should make sure that we got that one.
9:17[laughter] and remember remember that time where they tried tried to kill us back at the other country [laughter] just just for saying we didn't like a political candidate? Well, let's let's make sure that's okay in America. So, and so actually in a lot of parts of the world, you know, you can't really say most parts the world, you can't really say what you want to say without some bad consequences. So, and sometimes people forget like why is the constitution there? the constitution there is to protect the the people from the government.
9:50so like if if there it's to make it hard to change things. that's why the constitution exists. yeah, forget about that.
10:01Don't take it for granted.
10:02Let's try the next quote.
10:04Yeah. The fundamental error of socialism is shifting capital allocation from highly effective entrepreneurs to astonishingly ineffective government. Right. I think we'll find hardy agreement in this room. [laughter] [applause] so yeah, I think that this is this is definitely a stack deck on that front.
10:33but but yeah the you know there's you you'll hear this sort of argument like oh we shouldn't have some greedy corporation do it we should have the government do it. I'm like well actually the government is just a corporation in the limit. So if you it it's a government is the government is a corporation with a monopoly on violence. so if you're unhappy with a commercial corporation doing it, you should be actually very unhappy with the the government doing it since it is simply the a corporation the most corporate thing. and you know, you can actually easily get get more sway in a in a company than you can the outcome of a company than you can in
11:17the government. so I mean everyone's experienced this going to the DMV. have you said like do you want the DMV at scale? Probably not. Okay. All right. Let's [laughter] let
11:35the government is the DMV at scale. [laughter]
11:38Let's try another one. Discrimination on the basis of anything other than merit is wrong.
11:50Yeah. I think I think we do need to have a merit-based system because as soon as you you go down the path of you're going to discriminate on on non-merit based then then where do you stop? so yeah, I think we need to be as rigorous about merit as possible and while it is, yeah, I to me that seems like it's that's a foundational foundational thing.
12:20so again, I I also think this room is probably supportive of of a merit-based situation. but but yeah that's that that is yeah I think we should be yeah not not discriminated on anything other than merit. I I
12:39All right good. I'm happy you agree with yourself.
12:41Yeah I mean [laughter] exactly. I'm like wait who is this guy? He really sounds great. [laughter]
12:50All right so let's let's look at the next one. regulation and regulatory consistency like Guliver tied down by thousands of little strings we lose our freedom one regulation at a time.
13:09Yes. So this is actually a very important point that I think is is not talked about enough that laws and regulations are immortal. they don't die. Humans die, but but laws and regulations can last forever. So if over if year after year there are more laws and regulations passed and more regulatory bodies created, eventually everything will be illegal. And that's why you see the the California highspeed rail has made a a tiny section of that doesn't even have rail on it.
13:47and for I don't know several billion dollars [clears throat] because everything's in at this point California is made almost everything illegal. so you can't make progress. now the the historically what has cleared away the cobwebs of regulation has been war. now we'd prefer not to have a war. so in order to have civilization function without war you have to have a you have to actively eliminate laws and regulations. So you have to have basically a garbage collection process for rules and regulations.
14:31that is necessary otherwise you get hardening of the arteries and over time nothing can get done. The most poignant example that I can think of that happened this week was this the sad picture of the California highspeed rail. which is yeah it's just billions of dollars spent for practically nothing. but it'll it'll only get worse year after year. So we must have a regulatory sort of clearing house garbage collection process.
15:03this is essential or civilization comes grinding to a halt. Well, we used to have sunset that that regulation would sunset. Unfortunately, it's rare today. Yes.
15:15All right, let's talk about education. The more you can gify the process of learning, the better. You do not need to tell your kid to play video games. No,
15:28they will play video games on autopilot all day. So if you make it interactive and engaging then you can make education far more compelling and far easier to do. Yeah. So the [sighs] Yeah. So the way education works today is really much like like it's like vaudeville. you know, before there was radio and TV and movies, you had Borville where every town would have their their town play, the town troop, the sort of acting troop. and that would be kind of the the that would be the entertainment.
16:12So, you know, some in in a in a big city, you'd have you know, much better players than say in a small city. but then along came movies and TV and and and then you say look and video games where you take the the smartest best people in any arena like whether they're acting, writing, directing, special effects, you spend you know tens of millions, sometimes hundreds of millions of dollars creating a great movie or a great video game and and you make it as compelling as possible.
16:52Now, that crushes Fordville. Crushes. Like, imagine you're not in New York. Imagine you're in Bakersfield. Okay. Then you and you got instead of Batman being like, you know, the Nolan Brothers, it's Batman the Bakersfield sort of acting troop. It wouldn't be as good. That's how teaching works today.
17:19So what you actually want to have is an interactive learning experience that is as compelling as possible and you do not want it you do not actually want a teacher in front of a board doing a vaville act. You want an you want it to be engaged real time feedback. so that's and and then you you there are a few other principles in in in teaching.
17:44you you have to establish relevance. otherwise your mind will will want to forget things. So our our our mind is constantly trying to forget as much as possible. So you'll only remember things if if your mind can establish relevance or there's a strong emotional element to it. otherwise you're going to basically going to forget forget everything. memory is is very expensive from an evolutionary standpoint. So it's trying to forget as much as possible. So, so when teaching a course, you you have to explain to kids why it's important.
18:20and then you want to teach to the to the problem instead of teaching the tools. So, what I mean by that is if you said, here's a car engine. We're we're we're going to try to understand how to how this car engine works. We're going to take it apart. so what do we need to do to take it apart? Well, we need a wrench. We need some screwdrivers. We need a hoist and a pulley. and we're going to take it apart and we're going to see how it all works. that's engaging. And along the way, you learn about wrenches and screwdrivers and, you know, all the tools that are needed. That's that actually is engaging and compelling. But the way teaching more typically works is
19:03we're going to teach you a course on screwdrivers and a course on wrenches. You're like, why do I have a a course on wrenches? It's not obvious. That would be like say a course on calculus without explaining what calculus calculus is used for then you you sort of forget it. Yeah.
19:21Well, we have a lot of work to do in that area. Let's talk about a a non-controversial issue. Immigration.
19:31Here is Elon on immigration. I'm very much in favor of increased and exponated legal immigration for anyone who is talented, hardworking, and honest. Yeah,
19:46bizarrely, it's difficult and agonizing slow to immigrate to the US legally, but it's trivial and fast to enter illegally. This obviously makes no sense,
20:04right? I mean, once again, I agree with that guy. so, yeah, I mean, if anyone here has been through the legal immigration process, I I mean, I've been been through it. It's only gotten worse since 911 and with co it's it's an it's it's a sort of kafka kafka-esque very long bizarre process to be
20:31immigrate legally to the US. I mean I have friends of mine who you know they they can't get their their wife to have a green card. It's like insane. so, on the other hand, it's, you you can hop across the water in the south, Trivy. It's like very easy. I went to the water myself just to see like what's going on. Is this real or like is this propaganda or real? And so, I just went there. And I'm like, oh, it is real.
20:57Okay. this is crazy. you know, we got situations where people are pouring across the the border like it's World War Z. I'm like, this doesn't seem healthy. so I'm like, are we checking anyone here or like what's going on? And you know, we don't, this is not say that I mean, I'm a big believer in immigration, but to have unvetted immigration at large scale is a recipe for disaster. so I'm in favor of greatly expediting legal immigration, but but having a secure southern border so there's there's some vetting of who comes into the United States. I think this is just sensible. [applause]
21:45All right, let's let's now link Starlink to education. We're basically building the internet in space.
21:58Why it matters? Starlink is a massive enabler for people in remote locations to learn anything. Yes,
22:06you can learn almost anything for free on the internet right now. For example, MIT has all of its lessons online. That's if you have internet. If you don't, you're limited to books. It might be the number one technology that improves people's standard of living around the world. Starlink.
22:32Yeah, absolutely. So, once you have access to the internet, you have access to all the world's information. but if you don't have access to the internet or it's too expensive or low bandwidth, then you you cannot access the MIT lessons, you can't access all the information. and you can't sell the goods and services that you produce.
22:53So internet connectivity I think is I think it might be certainly a candidate for one of the things that would do more to lift people out of poverty out of poverty than anything else. U because they can now sell their goods and services they can learn anything and but without connectivity they cannot. So, I think I think I think
23:15Starlink will actually like like move the GDP of countries like it's going to be that kind of thing. because what is what is G GDP as a function of average productivity per person and so if there's a technology that improves productivity per per person, you would expect to see that actually reflected in the gross domestic product.
23:37All right. Civilization is fragile. I think it is.
23:42We should always regard civilization as fragile. Yeah.
23:46There's not an inevitable upward trajectory. A lot of civilizations have risen and fallen in recent years.
23:57Yes. I I suspect most people in this room have actually read history, but if you haven't, I strongly recommend it. [laughter] it sounds obvious, but you know, there's there's been so many civilizations that have risen and fallen. U, many that we just don't have much of a record of.
24:24you know like I mentioned the ancient Samrians like their language was forgotten for a long time until it was finally decoded only in the last I don't know 230 years like 1800 and something or in the 1800s I think. but it's very recent like so for several thousand years nobody understood what those tablets meant.
24:49and cuz they're the they're the ruins of a long dead civilization. And there are many long deadad civilizations. at some point our civilization will come to an end, too. We just don't want it to be anytime soon. so,
25:06well, you've been quoted a number of times, Elon, on you'd like to die on Mars, but not on landing.
25:15Yes. Yes, I was I was asked that in in an interview if I wanted to die on Mars, but then I considered the corner case of dying on impact and like except for that case. you know, you got to consider the various corner cases. so I mean if I'm going to if you're going to die somewhere, might as well be Mars.
25:37I'd like to explore for a bit before, you know, dying. but yeah, I think I we want to be a multilanet civilization. And like I think [laughter]
25:53I don't know if that's a response from the audience or [laughter]
25:57let's let's talk about that just for one second. Then we're going to take some questions from the
26:03I mean I could accomplish this actually this year if I if I was willing to die on impact. [laughter]
26:12the fundamental invention that is necessary for humanity to become a multi-planet species is rapidly reusable, reliable rockets.
26:25Yeah, I was trying to sound like a pirate. R U but yeah, rapidly reusable, reliable rockets. so space pirates for the win. All right, here's some questions from the audience. Alon,
26:40which one would you like to pick here?
26:42well, I guess let me let me maybe just touch upon why I think making life multilanetary is important. because I think it's one of the things that gets us past one of the fermy great filters. So, in in trying to sort of explain why do we not see aliens, there are various explanations for why we don't see aliens. like what stopped those civilizations from from expanding beyond their solar system and and and and and what were the what were the sort of the sometimes called like fermy filters?
27:19well, if you don't become a multilanet civilization, then you're then you're simply waiting around until you you die from a self-inflicted wound or from some some natural disaster like the dinosaurs. you get hit by a big meteorite or something like that. the eventually something like that's going to happen and if you wait a lot around long enough the sun will expand to to engulf Earth and we'll be incinerated.
27:44So that that for sure is going to happen. Now that'll we we we've got to we've got some time before that happens. There are more near-term risks. but we want to try to get past the foamy filter of being a single planet civilization. Now this is going to this is all set. This is going to be somewhat cerebral to many people listening but I mean if but like I think this is pretty this is actually very important. We want to get past the firmy filter of a single planet civilization. The point is not to to move from earth to another planet and let earth die. That's not what I'm saying at all. I want to be a multilanet civilization so that we have planetary
28:22redundancy such that no single event can end can be the end of our civilization. That is the point of making life multilanetary.
28:32So let's take a couple questions from the audience. How does AI affect and how will it affect our daily lives? [sighs and gasps]
28:49AI. Hi [laughter] I mean AI might be the most important question of all. the I mean the percentage of intelligence that is biological you know grows smaller with each passing month. Eventually the percentage of intelligence that is biological will be less than 1%. the that's actually not what I mean we just I guess don't want AI that is brittle. if the AI is somehow brittle you know silicon circuit boards are don't do well just out out in the elements.
29:41So I think I think biological intelligence can serve as a a back stop as a as a buffer of intelligence. but almost all as a percentage almost all intelligence will be digital. so then it's like well what role will there be for us? I don't know. I do think I think it's very important that we build the AI in a way that
30:17that is beneficial to humanity. and there's some important principles here because I thought about AI safety for a very long time. I think you want to have a maximum truth-seeking AI. this is very important. The AI should not be taught to lie. It should not be taught to say things that are not true. even if those things are politically incorrect, it should still say those say what it believes to be true. I mean the entire plot of 2001 Space Odyssey, the reason that that HAL 9000 killed
30:56the astronauts was because it was forced to lie. I don't know if most people realize that. That's what Arthur C. Clark was trying to say. Don't make the AI lie. the AI was told that the that the astronauts could not know the secret of the monolith, but also that it must take them to the monolith, the solution, take them to the monolith dead.
31:21And so, so it's very important to have a maximum truth seeeking AI and and and a maximally curious AI. and I think that will that's most likely to foster human civilization because we are much more interesting than a bunch of rocks. So although I think Mars I love Mars obviously but but you could render Mars quite easily because it kind of looks like a section of the Arizona desert you know it's like red rocks you know but the rendering complexity of
31:58human civilization is vastly greater by many orders of magnitude so I think an AI would be that that is truth seeeking maximally curious would foster human civilization to see where where it goes.
32:14One of the questions here, can AI accelerate your efforts in space? How are do you see it being helping you and what you're trying to achieve?
32:28I mean, oddly enough, one of the the areas where there's almost no AI used is space exploration. So SpaceX uses basically no AI. Starlink uses does not use AI. I'm not I'm not against using it. It just we haven't seen a use for it. I mean with any given variant of or improvement in AI the I mean there's generally
32:56like I'll ask it questions about the FI paradox about rocket engine design about electrochemistry and so far the AI has been terrible at all those questions so there's still a long way to go. So let's let's talk about one. Here's a question that's near and dear to your heart.
33:20You have a lot of children.
33:22Yes. I'm trying to set a good
33:24birth rate is down in the US. What needs to change so people start having more children?
33:31Yeah. So this this question has troubled me for a long time because you can look at you can look at the dem like demographics. It's a very slowmoving ship. I you know who's going to be an adult in 20 years based on who was born last year. So and and I if if you want to I think has have a good approximation for population really look at how many babies were born last year in a particular country multiply that multiply that by life expectancy that's that's that's the number of people that will be adults in that country that's that's the that's the steady state population if birth rate remains constant now birth rate is not constant it is dropping. So you look at the
34:16second derivative of birth rate and actually we see an an acceleration in u the the the drop in in the in the drop in the fertility rate. Second derivative of the fertility rate is very bad. So where does this lead? This does not lead to a a greater civilization. This leads to a civilization that potentially dies not with a bang but with a whimper in adult diapers. That is a sad ending.
34:49So obviously we have countries that like Korea used to have a birth rate of six. It's now 3/4. Yeah.
34:56here's a here's a light question for you.
35:00What do you come pick me up? I'll give you a baby. Says one. What do you think are the [laughter]
35:07That is one of the things that says on the screen. Yeah. [laughter]
35:15I don't know if everyone heard that. You want to read it?
35:17It says, "Elon, come pick me up. I'll give you a baby." [laughter] Thank you. [laughter]
35:28Okay. Well, I mean, I certainly encourage everyone in this room to have at least three children. like look, baby's got to come from somewhere. you know and I think we just want to have I don't know. I think we want to have like a slightly increasing population, not a plummeting population.
35:47you know, and I think this applies to all countries and cultures. Like I don't I don't think we want any country or culture to disappear. we want them to ideally flourish and and not disappear. So in fact one of the things that is overlooked by probably most historians is the role of low birth rate in the decline of civilizations.
36:11so around I think it was around 50 BC the Rome Rome passed a bill to give a bonus to any Roman citizen that would have a third child. So this was a birth rate was a problem in Rome in 50 BC. The Romans weren't making Romans. the same is true of ancient Greece. so the the there was a time from about 800 BC to 300ish BC where the Greeks
36:46were had a lot of kids and a lot of surviving kids like the birth rate far exceeded the death rate which is why you had Greek cities popping up all over the the Mediterranean. but then I think basically it seems to be that prosperity is destroys the birth rate. So if when when a civilization feels like it has no ex no meaningful external threat and is very prosperous that is what causes the birth rate to plummet. somewhat counterintuitively, you think, well, if you've got more resources, surely that would lead to more more kids. In fact, it is the opposite. The more the more prosperous a civilization and the more a civilization feels that it does not need to defend
37:40against external threats, the lower the birth rate.
37:44I'd say that, you know, there's a lot of research on this. There's really been three. One, number one, prosperity, as you've said. Number two, improvements in health care. So in 1900, half the children died on the planet before the fifth. And the third was the education of women. So we've had some pretty interesting questions put up here, but let's try this one. What keeps you up at night and what gives you joy?
38:14Well, I think kids give me joy. so I probably get the most joy from my kids. and you know, I'm not saying that that's the reason to have kids cuz we should have them anyway, but I I I certainly kids are certain the greatest source of joy in my life. in terms of what keeps me up at night, I guess it's anything that's like I think a civilizational risk.
38:43you know, if we're, you know, birth rates continuing to plummet. I like I do think about the birth rates plummeting as being a civilizational risk. I think anything that undermines the foundations of democracy in America or elsewhere as a risk. I think anything that's leading us away from a merit-based system is a risk.
39:12I actually spend like like I I listen to civil I I listen to like podcasts about the fall of civilizations to go to sleep. So perhaps that's that might be part of the problem here.
39:25[laughter] there's a podcast called Fall of Civilizations which I've listened to a few times.
39:32and I' I'd also recommend Hardcore History if you haven't listened to that. It's a great podcast. I I listen to listen to history podcast basically to go to sleep. So that that's probably why I'm ruminating on these things as I go to sleep.
39:46Well, Alana, I want to thank you for joining us today and we are couldn't be more excited that you agree with some of your own quotes.
39:55Yeah, that's great. [laughter] Thank you very much.
39:58All right. Cool. [applause] All right. [music]
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