埃隆·马斯克谈战争、人性与中东和平

以下是与埃隆·马斯克的对话，这是他第四次上《莱克斯·弗里德曼播客》。我以为你会把它弹完。这是电影史上最伟大的主题之一。是的，很棒。所以我刚才在想罗马帝国，就像人们常做的那样。有个梗说所有男人每天至少想一次罗马帝国，而一半的人困惑这是不是真的。但更严肃地说，想想当今世界的战争，如你所知，战争和军事征服一直是罗马社会和文化的重要组成部分，我认为这也是人类历史上大多数帝国和王朝的重要组成部分。是的，它们通常源于征服。有些像奥匈帝国那样，有很多巧妙的联姻，但根本上还是征服的引擎，他们崇尚战争中的卓越。许多领袖都是优秀的将军，是的，诸如此类。所以一个大问题：Grok 批准了吗？我问了这个问题，是个好问题。测试过，Grok 批准了，至少在趣味模式下。你认为战争在多大程度上是人性的一部分，而不是人类社会结构的结果？我问这个是因为你不知为何有争议地成了和平的支持者。

The following is a conversation with Elon Musk, his fourth time on The Lex Fridman podcast. I thought you were going to finish it. It's one of the greatest themes in all film history. Yeah, that's great. So I was just thinking about the Roman Empire, as one does. There's that whole meme where all guys are thinking about the Roman Empire at least once a day, and half the population is confused whether it's true or not. But more seriously, thinking about the wars going on in the world today, and as you know, war and military conquest has been a big part of Roman society and culture, and I think it has been a big part of most empires and dynasties throughout human history. So yeah, they usually came as a result of conquest. I mean, there's some like the Austro-Hungarian Empire where there was just a lot of clever marriages, but fundamentally there's an engine of conquest, and they celebrate excellence in warfare. Many of the leaders were excellent generals, yeah, that kind of thing. So big picture question: Grok approved? I asked this is a good question to ask. Tested Grok approved, at least on fun mode. To what degree do you think war is part of human nature versus a consequence of how human societies are structured? I asked this as you have somehow controversially been a proponent of peace.

我总体上支持和平。我的意思是，在我看来，无知才是真正需要对抗的敌人。那才是真正的难点，而不是与其他人战斗。但所有生物都会争斗。我的意思是，丛林——人们认为大自然也许是某种和平的东西，但实际上并非如此。有个挺搞笑的 W. W. H. 的东西，他说在丛林里，基本上四面八方都是谋杀和死亡。我是说，丛林里的植物和动物每天都在每分每秒地试图互相残杀。所以在这方面我们并不特别。

I'm generally a proponent of peace. I mean, ignorance is perhaps in my view the real enemy to be countered. That's the real hard part, not fighting other humans. But all creatures fight. I mean, the jungle—you look at people think of nature as perhaps some sort of peaceful thing, but in fact it is not. There's some quite funny W. W. H. thing where he's like in the jungle, saying that it's like basically just murder and death in every direction. I mean, the plants and animals in the jungle are constantly trying to kill each other every single day, every minute. So it's not like we're unusual in that respect.

嗯，这里有个相关的问题：更高的智能是否带来对这些暴力本能的更强控制。

Well, there's a relevant question here: whether with greater intelligence comes greater control over these base instincts for violence.

是的，我们比黑猩猩有更强的控制暴力本能的能力。事实上，如果你看看黑猩猩社会，它并不友好。我的意思是，倭黑猩猩是个例外，但黑猩猩社会充满了暴力，坦率说非常可怕。那是边缘系统在起作用。你不想站在黑猩猩的对立面；它会吃掉你的脸，扯掉你的蛋。是的，基本上没有限制或道德。罗马人讲“正义战争”；黑猩猩社会没有正义战争。那是不择手段的战争和统治。黑猩猩社会是人类社会的原始版本。它们根本不爱好和平。有极端的暴力。然后偶尔有人看了太多迪士尼电影，决定养一只黑猩猩当宠物，然后那只黑猩猩就吃掉他们的脸，或者扯掉他们的蛋，或者咬掉他们的手指。这种事发生过好几次。

Yes, we have much more ability to control our limbic instinct for violence than say a chimpanzee. And in fact, if one looks at chimpanzee society, it is not friendly. I mean, the bonobos are an exception, but chimpanzee society is filled with violence and it's quite horrific frankly. That's the limbic system in action. You don't want to be on the wrong side of a chimpanzee; he'll eat your face off and tear your nuts off. Yeah, basically there's no limits or ethics. The Romans said 'just war'; there's no just war in chimpanzee societies. It's war and dominance by any means necessary. Chimpanzee society is like a primitive version of human society. They're not peace-loving at all. There's extreme violence. And then once in a while, somebody who's watched too many Disney movies decides to raise a chimpanzee as a pet, and then that eats their face or tears their nuts off or chews their fingers off. That kind of thing. It's happened several times.

扯掉你的蛋是一种有趣的互动策略。这发生在人身上过。很不幸。我想这是确保另一只黑猩猩不贡献基因库的一种方式。嗯，从武术角度看，这是一种迷人的策略。我想知道哪种武术教这个。可以安全地说，如果有人抓住了你的蛋，并且可以选择扯掉它们，你会顺从他们想要的任何东西。是的，可以这么说。所以就像我说的，不知为何有争议地，你在 Twitter 和 X 上一直是和平的支持者。那么让我问问你关于当今的战争，以及和平之路可能是什么。你希望当前以色列和加沙的战争如何结束？你看到什么路径可以长期最小化那个地区的人类苦难？

Ripping your nuts off is an interesting strategy for interaction. It's happened to people. It's unfortunate. That's I guess one way to ensure that the other chimp doesn't contribute to the gene pool. Well, from a martial arts perspective, it's a fascinating strategy. I wonder which of the martial arts teaches that. I think it's safe to say if somebody's got your nuts in their hand and has the option of ripping them off, you will be amenable to whatever they want. Yeah, so to say. So like I said, somehow controversially, you've been a proponent of peace on Twitter, on X. So let me ask you about the wars going on today and to see what the path to peace could be. How do you hope the current war in Israel and Gaza comes to an end? What path do you see that can minimize human suffering in the long term in that part of the world?

嗯，我认为那个地区绝对是——如果你在字典里查“没有简单答案”，它会是一张中东的图片，尤其以色列。所以没有简单答案。严格来说我的看法是，哈马斯的目标是激怒以色列过度反应。他们显然不指望军事胜利，而是想尽可能犯下最严重的暴行，以引发以色列最激进的回应，然后利用那种激进回应来动员全世界的穆斯林支持加沙和巴勒斯坦的事业，他们成功了。所以这里反直觉的是，我认为应该做的事情，尽管非常困难，是建议以色列进行尽可能引人注目的善举。一切会挫败哈马斯目标的事情。在某种意义上，在符合地缘政治逻辑的程度上，实行“转过另一边脸”。不完全是转过另一边脸，因为我确实认为以色列找到哈马斯成员并杀死或监禁他们是合适的。必须做点什么，否则他们会不断再来。但除此之外，他们需要尽一切可能。有讨论建立移动医院；我建议这么做。确保有食物、水、医疗必需品，并且做得过头一点，非常透明，这样人们就不能说这是个诡计。放个网络摄像头，24/7。部署善举。明确无误的引人注目的善举，意思是它们不能以某种方式——因为哈马斯会说这是个诡计，所以你必须反驳。这最终会对抗该地区更广泛的仇恨力量。是的。我不确定是谁说的，这是个伪托的谚语，但“以眼还眼使所有人盲目”。在那个地区，他们真的相信以眼还眼那一套。但如果你不打算直接对整个人群实施种族灭绝——这显然对任何人都是不可接受的——那么你基本上会留下很多活着的人，他们随后会仇恨以色列。所以真正的问题是：每杀死一个哈马斯成员，你创造了多少个？如果你创造的比杀死的多，你就没有成功。这就是那里的真实情况。

Well, I think that part of the world is definitely like if you look up 'there is no easy answer' in the dictionary, it'll be a picture of the Middle East, in Israel especially. So there is no easy answer. What is strictly my opinion is that the goal of Hamas was to provoke an overreaction from Israel. They obviously did not expect to have a military victory, but they really wanted to commit the worst atrocities they could in order to provoke the most aggressive response possible from Israel, and then leverage that aggressive response to rally Muslims worldwide for the cause of Gaza and Palestine, which they have succeeded in doing. So the counterintuitive thing here is that I think the thing that should be done, even though it is very difficult, is that I would recommend that Israel engage in the most conspicuous acts of kindness possible. Everything that would thwart the goal of Hamas. In some sense, to the degree that makes sense in geopolitics, turn the other cheek implemented. It's not exactly turn the other cheek, because I do think it is appropriate for Israel to find the Hamas members and either kill them or incarcerate them. Something has to be done because they're just going to keep coming otherwise. But in addition to that, they need to do whatever they can. There's some talk of establishing a mobile hospital; I'd recommend doing that. Just making sure that there is food, water, medical necessities, and just be over the top about it and be very transparent, so that people can't claim it's a trick. Just put a webcam on the thing, 24/7. Deploy acts of kindness. Conspicuous acts of kindness that are unequivocal, meaning they can't be somehow—because Hamas will then say it's a trick, so you have to counter that. This ultimately fights the broader force of hatred in the region. Yes. And I'm not sure who said it, it's an apocryphal saying, but 'an eye for an eye makes everyone blind.' Now in that neck of the woods, they really believe in the whole eye for an eye thing. But if you're not going to just outright commit genocide against an entire people, which obviously would not be acceptable to anyone, then you're going to leave basically a lot of people alive who subsequently hate Israel. So really the question is: for every Hamas member that you kill, how many did you create? And if you create more than you killed, you've not succeeded. That's the real situation there.

你在加沙杀了别人的孩子，你就制造了至少几个哈马斯成员，他们会不惜牺牲生命去杀一个以色列人。情况就是这样。但这是最可能引发争议的话题之一。但如果最终目标是某种长期和平，就必须从长远来看，随着时间的推移，是产生了更多还是更少的恐怖分子？

You kill somebody's child in Gaza, you've made at least a few Hamas members who will die just to kill an Israeli. That's the situation. But this is one of the most contentious subjects one could possibly discuss. But if the ultimate goal is some sort of long-term peace, one has to look at this from the standpoint of over time, are there more or fewer terrorists being created?

让我就停留在战争这个话题上。可以说，战争一直存在，也永远会存在。

Let me just linger on war. Well, it's safe to say wars always existed and always will exist.

永远会存在？我希望不会。

Always will exist? I hope not.

我认为战争永远存在。问题只是战争的多少，以及战争的规模和范围。但想象未来没有任何战争，我认为是非常不可能的结果。

I think there will always be war. The question is just how much war, and the scope and scale of war. But to imagine that there would not be any war in the future, I think would be a very unlikely outcome.

你提到过《文明》系列；那里也有战争。是的，一场巨大的战争。第一本书就以一场巨大的银河战争开场，数万亿人死亡。但它仍然保护着那些繁荣的角落。不知怎的，你可以有银河战争，同时仍有繁荣的角落。

You talked about the Culture series; there's war even there. Yes, a giant war. The first book starts off with a gigantic galactic war where trillions die. But it still protects these pockets of flourishing. Somehow you can have galactic war and still have pockets of flourishing.

是的，我的意思是，这是苦难规模与繁荣规模之间的竞赛。

Yeah, I mean, it's a race between the scale of suffering and the scale of flourishing.

很多人似乎正在利用这场悲剧来敲响战鼓，滋养军工复合体。你担心这个吗？那些支持局势升级的人，怎样才能阻止他们？

A lot of people seem to be using this tragedy to beat the drums of war and feed the military-industrial complex. Do you worry about this? The people who are rooting for escalation, and how can it be stopped?

让我担心的一件事是，今天活着的人中，很少有人真正切身理解战争的恐怖，至少在美国是这样。我的意思是，显然在乌克兰和俄罗斯前线的人理解战争有多可怕。但西方有多少人理解呢？我的祖父参加过二战。他受到了严重的创伤。他在北非和意大利待了将近六年。他所有的朋友都在他面前被杀，他本来也会死，但碰巧他们随机进行了一些智商测试之类的，他得分很高。他不是军官；他是下士或中士，因为他没读完高中。他不得不辍学，因为他父亲去世了，他得工作养活兄弟姐妹。所以因为他没高中毕业，他没有资格成为军官。他基本上被归为炮灰。但后来他们随机给了他这个测试，他被调到了伦敦的英国情报机构。在那里他遇到了我的祖母。但他有创伤后应激障碍，非常严重。他就是不说话。如果你试图和他说话，他就叫你闭嘴。他赢得了一堆勋章，从未吹嘘过一次，甚至没有暗示过。我是因为他的军事记录在网上才知道的。他会说，‘你绝对不想再来一次。’但现在有多少人记得二战？不多了。同样的情况可能也适用于核战争的威胁。

One of the things that does concern me is that there are very few people alive today who actually viscerally understand the horrors of war, at least in the US. I mean, obviously the people on the front lines in Ukraine and Russia understand just how terrible war is. But how many people in the West understand it? My grandfather was in World War II. He was severely traumatized. He was there for almost six years in North Africa and Italy. All his friends were killed in front of him, and he would have died too, except they randomly gave some IQ tests or something, and he scored very high. He was not an officer; he was a corporal or a sergeant because he didn't finish high school. He had to drop out because his dad died, and he had to work to support his siblings. So because he didn't graduate high school, he was not eligible for the officer corps. He got put into the cannon fodder category basically. But then randomly they gave him this test, and he was transferred to British intelligence in London. That's where he met my grandmother. But he had PTSD, next level. He just didn't talk. If you tried talking to him, he'd just tell you to shut up. He won a bunch of medals, never bragged about it once, not even hinted. I found out because his military records are online. He would say, 'No way in hell do you want to do that again.' But how many people now remember World War II? Not many. And the same perhaps applies to the threat of nuclear war.

是的，我的意思是，有足够多的核弹瞄准美国，足以让放射性瓦砾多次弹跳。现在有两场大战正在进行。你谈了很多关于 AGI 的威胁，但现在我们坐在这里，冲突如此激烈，你担心核战争吗？

Yeah, I mean, there are enough nuclear bombs pointed at the United States to make the radioactive rubble bounce many times. There are two major wars going on right now. You talked about the threat of AGI quite a bit, but now as we sit here with the intensity of conflict going on, do you worry about nuclear war?

我认为我们不应该低估核战争的可能性。这是一种文明威胁。现在，我可能错了，但我认为当前核战争的概率相当低。但有很多核弹瞄准我们，我们也有很多核弹瞄准别人。它们还在那里。没有人放下武器。导弹还在发射井里，而拥有核武器的领导人似乎没有在互相交谈。

I think we shouldn't discount the possibility of nuclear war. It is a civilizational threat. Right now, I could be wrong, but I think the current probability of nuclear war is quite low. But there are a lot of nukes pointed at us, and we have a lot of nukes pointed at other people. They're still there. Nobody has put their guns away. The missiles are still in the silos, and the leaders don't seem to be the ones with the nukes talking to each other.

有些战争在局部层面是悲惨和艰难的，而有些战争则是文明终结或具有这种潜力。显然，全球热核战争极有可能终结文明，也许是永久性的，但肯定会严重伤害并阻碍人类进步，回到石器时代或类似情况。我不知道，非常糟糕。可能在那之后科学家和工程师也不会太受欢迎。他们会说，‘是你们把我们搞成这样的。’所以总的来说，我们显然应该优先考虑文明风险，而不是那些在局部层面痛苦悲惨但非文明性的问题。

There are wars which are tragic and difficult on a local basis, and then there are wars which are civilization-ending or have that potential. Obviously global thermonuclear warfare has high potential to end civilization, perhaps permanently, but certainly to severely wound and set back human progress to the Stone Age or something. I don't know, pretty bad. Probably scientists and engineers won't be super popular after that either. They'll be like, 'You got us into this mess.' So generally, we obviously want to prioritize civilizational risks over things that are painful and tragic on a local level but not civilizational.

你希望乌克兰战争如何结束？再次最小化那里人类苦难的路径是什么？

How do you hope the war in Ukraine comes to an end, and what's the path once again to minimizing human suffering there?

嗯，我认为很可能发生的情况，实际上几乎就是现在这样，是停火或休战将沿着非常接近当前战线的地方发生。但你知道，现在的情况是，无论哪一方发动进攻，伤亡率都会是防守方的数倍，因为你有纵深防御：雷区、战壕、反坦克防御。没有人拥有空中优势，因为防空导弹远比飞机厉害，而且数量多得多。所以双方都没有空中优势。坦克基本上是死亡陷阱，移动缓慢，而且对反坦克武器不免疫。所以你实际上只有远程火炮和步兵战壕。这又是一战，只是多了无人机。无人机让远程火炮更加精确和有效，从而更高效地屠杀双方人员。所以任何一方都不想尝试推进，因为死亡概率极高。要克服纵深防御、战壕和雷区，你真的需要显著的局部数量优势，最好结合飞机快速攻击、集中数量的坦克和大量人员。这是你突破防线然后不被援军立刻赶回去的唯一方法。我真的建议人们详细阅读一战的历史。那很残酷。死亡人数之多令人难以置信，而且几乎无法想象一个不……的结局。

Well, I think what is likely to happen, which is really pretty much the way it is, is that something very close to the current lines will be how a ceasefire or truce happens. But you know, you have a situation right now where whoever goes on the offensive will suffer casualties at several times the rate of whoever is on the defense, because you've got defense in depth: minefields, trenches, anti-tank defenses. Nobody has air superiority because anti-aircraft missiles are really far better than aircraft, and there are far more of them. So neither side has air superiority. Tanks are basically death traps, slow-moving and not immune to anti-tank weapons. So you really just have long-range artillery and infantry trenches. It's World War I all over again, with drones. Drones make the long-range artillery that much more accurate and better, so more efficient at murdering people on both sides. So you don't want to be trying to advance from either side because the probability of dying is incredibly high. In order to overcome defense in depth, trenches, and minefields, you really need significant local superiority in numbers, ideally combined with a fast attack with aircraft, a concentrated number of tanks, and a lot of people. That's the only way you're going to punch through a line and then not have reinforcements kick you right out again. I really recommend people read about World War I warfare in detail. It's rough. The sheer number of people that died was mind-boggling, and it's almost impossible to imagine the end of it that doesn't...

看起来几乎和开始时一模一样，土地归属等等都没变，但另一边却是大量的人类苦难、死亡和基础设施的破坏。

It looks like almost exactly like the beginning in terms of who what land belongs to who and so on, but on the other side of a lot of human suffering, death, and destruction of infrastructure.

是的。我一年前提出某种休战或和平的原因，正是因为我几乎准确预测了会发生什么——大量的人死去，而领土基本没有变化。这是乌克兰和俄罗斯青年精英的损失。我们应该对俄罗斯男孩和乌克兰男孩都抱有同情，因为俄罗斯男孩并非自愿上前线，他们不得不去。很多儿子再也回不到父母身边。我认为他们大多数并不真正仇恨对方。就像一战时的那句话：互不相识的年轻男孩，替彼此认识的老男人互相残杀。这有什么意义呢？

Yes. The reason I proposed some sort of truce or peace a year ago was because I predicted pretty much exactly what would happen, which is a lot of people dying for basically almost no changes in land. And this is the loss of the flower of Ukrainian and Russian youth. We should have some sympathy for the Russian boys as well as the Ukrainian boys, because the Russian boys didn't ask to be on the front line; they have to be. So there are a lot of sons not coming back to their parents. I think most of them don't really hate the other side. It's like the saying from World War I: young boys who don't know each other killing each other on behalf of old men that do know each other. What's the point of that?

泽连斯基过去说过他对直接与普京对话不感兴趣。你认为他应该谈吗？

Volodymyr Zelenskyy has said in the past he's not interested in talking to Putin directly. Do you think he should?

是的，面对面坐下来，领袖对领袖，谈判和平。听着，我只建议：不要把乌克兰的青年精英送去战壕里送死。无论他是否与普京对话，都不要这样做。任何发动进攻的一方都将损失大量人员，历史不会善待他们。

Yeah, sit down man-to-man, leader to leader, and negotiate peace. Look, I think I would just recommend: do not send the flower of Ukrainian youth to die in trenches. Whether he talks to Putin or not, just don't do that. Whoever goes on the offensive will lose massive numbers of people, and history will not look kindly upon them.

你曾坦诚地谈到，如果找不到外交解决方案，比如在台湾问题和一中政策上，美中长期来看可能发生战争。我们如何避免这两个超级大国冲突的轨迹？

You've spoken honestly about the possibility of war between the US and China in the long term if no diplomatic solution is found, for example on the question of Taiwan and the One China policy. How do we avoid the trajectory where these two superpowers clash?

嗯，值得一读那本名字难念的《修昔底德陷阱》。我热爱战争史，倒背《孙子兵法》和《战争艺术》。几乎没有我没读过的战役，我试图弄清楚每场胜利的真正原因，而不是某一方声称的原因。无论是胜利还是战争的起因，整个事情。雅典和斯巴达就是经典案例。希腊人的特点是他们记录了很多东西，他们热爱写作。世界上很多地方发生了很多有趣的事，但人们没有记录下来，所以我们不知道发生了什么。或者他们记录得不详细，只会说‘我们打了一仗，赢了’。能多说点吗？希腊人真的写了很多，他们口才很好，就是热爱写作。我们保存了其中一些文字，所以知道斯巴达和雅典联盟之间伯罗奔尼撒战争的起因。我们知道他们预见到了。斯巴达人不怎么写，他们天性不善言辞。但雅典人和其他希腊人写了很多。他们说：‘斯巴达曾是希腊的领袖，但雅典一年比一年强大。’大家都认为雅典和斯巴达之间的冲突不可避免。‘我们如何避免？’他们做不到。他们实际上预见到了，但仍然无法避免。所以到了某个点，如果一个群体、文明或国家超过了另一个……美国自 1890 年左右以来一直是经济上最大的玩家。美国成为世界上最强大的经济引擎的时间比任何在世的人都要长。战争的基础是经济。现在中国的情况是，其经济规模可能是美国的两到三倍。想象一下，你是有史以来最大的玩家，突然来了一个比你大一倍的孩子。我们预见到了。怎么可能阻止呢？

Well, it's worth reading that book on the difficult-to-pronounce Thucydides Trap. I love war history. I like Sun Tzu and Art of War backwards. There's hardly a battle I haven't read about, trying to figure out what really was the cause of victory in any particular case, as opposed to what one side or another claimed. Both the victory and what sparked the war. The whole thing. So Athens and Sparta is a classic case. The thing about the Greeks is they really wrote down a lot of stuff; they loved writing. There are lots of interesting things that happened in many parts of the world, but people just didn't write down, so we don't know what happened. Or they didn't write in detail; they'd just say, 'We had a battle and we won.' Like, can you add a bit more? The Greeks really wrote a lot; they were very articulate and just loved writing. We have a bunch of that writing preserved, so we know what led up to the Peloponnesian War between the Spartan and Athenian alliances. We know that they saw it coming. The Spartans didn't write; they weren't very verbose by nature. But the Athenians and other Greeks wrote a lot. They were like, 'Sparta was really the leader of Greece, but Athens grew stronger with each passing year.' Everyone thought it was inevitable that there would be a clash between Athens and Sparta. 'How do we avoid that?' And they couldn't. They actually saw it coming and still could not avoid it. So at some point, if one group, civilization, or country exceeds another... The United States has been the biggest kid on the block since around 1890 from an economic standpoint. The US has been the most powerful economic engine in the world longer than anyone's been alive. The foundation of war is economics. Now we have a situation with China where the economy is likely to be two, perhaps three times larger than that of the US. So imagine you're the biggest kid on the block for as long as anyone can remember, and suddenly a kid comes along who's twice your size. We see it coming. How is it possible to stop?

有没有一些文化交融和理解？双方似乎存在巨大的文化理解鸿沟。你是一个有趣的案例，因为你是美国人，在美国做了很多制造业，但你也与中国合作。我在中国待过很长时间，多次见过中国领导人。也许一个好的问题是：关于中国，有哪些人们不了解的、文化上积极的东西？你学到了哪些关于中国人的有趣事情？

Is there some intermixing of cultures, understanding? There does seem to be a giant cultural gap in understanding of each other. You're an interesting case study because you are an American, you've done a lot of manufacturing here in the US, but you also work with China. I've spent a lot of time in China and met with the leadership many times. Maybe a good question to ask is: what are some things about China that people don't understand, positive just in the culture? What interesting things have you learned about the Chinese?

嗯，中国真正聪明勤奋的人数量惊人。在我看来，那里比这里多得多。而且他们精力充沛。近年来中国的建筑比美国令人印象深刻得多。火车站、高楼、高铁，一切都比我们美国的好得多。我建议人们去上海和北京看看那些建筑，然后坐火车从北京到西安看兵马俑。中国有令人难以置信的历史，非常悠久。我认为，从文字使用的角度来看，它是最古老的语言之一，也许是最古老的。而且中国人确实记录了很多东西。现在，历史上中国除了极少数例外，一直专注于内部。他们不好奇。他们互相争斗，内战很多。在三国战争期间，我相信他们损失了大约 70%的人口。所以他们有过残酷的内战，相比之下美国内战都显得规模小了。我认为重要的是要认识到中国不是铁板一块。我们有点把中国看作一个单一的整体，有统一的思想，但这绝对不是事实。根据我的观察，而且我认为大多数了解中国的人都会同意，中国人思考中国内部事务的时间是思考外部事务的十倍。所以他们的考虑 90%是内部的。

Well, the sheer number of really smart, hardworking people in China is incredible. There are far more of them there than there are here, in my opinion. And they've got a lot of energy. The architecture in China in recent years is far more impressive than in the US. The train stations, the buildings, the high-speed rail, everything is really far more impressive than what we have in the US. I recommend somebody just go to Shanghai and Beijing, look at the buildings, and take the train from Beijing to Xi'an where you have the Terracotta Warriors. China has incredible history, a very long history. I think arguably, in terms of the use of language from a written standpoint, it's one of the oldest, perhaps the oldest written language. And Chinese people did write things down. Now, China historically has always been, with rare exception, internally focused. They have not been inquisitive. They fought each other; there have been many civil wars. In the Three Kingdoms War, I believe they lost about 70% of their population. So they've had brutal internal wars that make the US Civil War look small by comparison. I think it's important to appreciate that China is not monolithic. We sort of think of China as a single entity of one mind, and this is definitely not the case. From what I've seen, and I think most people who understand China would agree, people in China think about China ten times more than they think about anything outside of China. So 90% of their consideration is internal.

当谈到超级大国之间未来和平合作时，这难道不是一件非常积极的事情吗？当你向内看，专注于提升自己，而不是通过军事力量去‘提升’别人。

Isn't that a really positive thing when you're talking about collaboration in a future peace between superpowers? When you're inward-facing, focusing on improving yourself versus focusing on improving others through military might.

好消息是：中国的历史……

The good news: the history of China...

这表明中国并不具有侵略性，意思是他们不会去入侵一大堆国家。不过他们确实有很强的立场，这很好。因为很多非常强大的国家都具有侵略性。美国也是少数不具有侵略性的国家之一。比如二战后，美国本可以基本上接管全世界。我们有核武器，其他国家没有。我们甚至不用损失士兵。你想要哪个国家？美国本可以接管一切，但它没有。美国实际上帮助重建了其他国家，比如帮助重建了欧洲，帮助重建了日本。这是非常不寻常的行为，几乎是史无前例的。美国做了像柏林空运这样引人注目的善举。而且我想，人们总是说，美国做过坏事。当然美国做过坏事，但需要看整体记录。一般来说，一个检验标准是你如何对待战俘。或者说，无意冒犯俄罗斯人，但假设你在 1945 年的德国，一边是苏联军队，另一边是法、英、美军队，你愿意向谁投降？没有哪个国家是完美道德的，但我建议成为美国的战俘。那会是我的强烈选择。事实上，冯·布劳恩，聪明人，就说我们必须被美国人俘虏。实际上，党卫军曾接到命令处决冯·布劳恩和所有德国火箭工程师。他们侥幸逃脱。他们说要去树林里散步，在隆冬时节没有穿外套就离开了，然后没有食物、没有外套、没有水，拼命向西跑。幸运的是，他兄弟找到了一辆自行车之类的，然后他尽可能快地骑向西，找到了一支美军巡逻队。所以，这是判断道德的一种方式：你想成为哪国的战俘？哪里都不好玩，但有些地方比别的地方糟糕得多。所以，美国虽然远非完美，但总体上是一股仁慈的力量。我们应该总是自我批评，努力变得更好。但任何有半个脑子的人都知道这一点。所以我认为在这方面，中国和美国是相似的。两个国家都没有显著的侵略性。所以这大概是一个共同原则。

Suggests that China is not inquisitive, meaning they're not going to go out and invade a whole bunch of countries. Now they do feel very strongly, you know. So that's good, I mean because a lot of very powerful countries have been inquisitive. The US is one of the also one of the rare cases that has not been inquisitive. Like after World War II, the US could have basically taken over the world and any country. Like we got nukes, nobody else got nukes. We don't even have to lose soldiers. Which country do you want? And the United States could have taken over everything. And it didn't. The United States actually helped rebuild countries, so it helped rebuild Europe, you know, helped rebuild Japan. This is very unusual behavior, almost unprecedented. You know, the US did conspicuous acts of kindness like the Berlin airlift. And I think you know, there it's always like, well America's done bad things. Well of course America's done bad things, but one needs to look at the whole track record. And generally, one sort of test would be how do you treat your prisoners of war. Or let's say, no offense to the Russians, but let's say you're in Germany in 1945, you got the Russian army coming one side, you got the French, British and American armies coming the other side. Who would you like to surrender to? No country is perfectly moral, but I recommend being a POW with the Americans. That would be my choice very strongly. And in fact, Von Braun, smart guy, was like, we've got to be captured by the Americans. And in fact, the SS was under orders to execute Von Braun and all of the German rocket engineers. And they narrowly escaped. They said they were going out for a walk in the woods. They left in the middle of winter with no coats, and they ran with no food, no coats, no water, and just ran like hell, and ran west. And by some luck, I think his brother found a bicycle or something, and then just cycled west as fast as he could and found a US patrol. So anyway, that's one way you can tell morality: who do you want to be a POW? It's not fun anywhere, but some places are much worse than others. So anyway, America has been, while far from perfect, generally a benevolent force. And we should always be self-critical and try to be better. But anyone with half a brain knows that. So I think in this way, China and the United States are similar. Neither country has been inquisitive in a significant way. So that's like a shared principle, I guess.

现在中国对台湾问题立场非常坚定。他们长期以来一直非常明确。从他们的角度看，这就像某个州不在了，比如夏威夷之类的，但比夏威夷更重要。夏威夷对我们来说也很重要。所以他们将其视为中国不可分割的一部分。福尔摩沙岛，不是台湾，它不属于中国但应该属于。而它之所以还没有，是因为美国太平洋舰队。随着经济实力和军事实力的增长，他们明确表示的利益将会实现。是的，中国已经非常明确，他们将和平或军事地统一台湾，但一定会统一。从他们的角度看，这是 100%确定的。

Now China does feel very strongly about Taiwan. They've been very clear about that for a long time. From their standpoint, it would be like one of the states is not there, like Hawaii or something like that, but more significant than Hawaii. Hawaii is pretty significant for us. So they view it as really a fundamental part of China. The island of Formosa, not Taiwan, that is not part of China but should be. And the only reason it hasn't been is because of the US Pacific Fleet. And as their economic power grows and as their military power grows, the thing that they are clearly saying is their interest will clearly be materialized. Yes, China has been very clear that they will incorporate Taiwan peacefully or militarily, but that they will incorporate it. From their standpoint, it's 100% likely.

你提到的引人注目的善举作为一种地缘政治策略，听起来几乎有些天真，但我敢说这可能是前进的道路，是避免大多数战争的方法。就像你说的，听起来天真但很聪明，如果你相信人性本善。似乎引人注目的善举可以在相关国家民众中产生共鸣，并缓和局势。

Something you said about conspicuous acts of kindness as a geopolitical policy, it almost seems naive, but I'd venture to say that this is probably the path forward, how you avoid most wars. Just as you say, it sounds naive but it's kind of brilliant if you believe in the goodness underlying most of human nature. It just seems like conspicuous acts of kindness can reverberate through the populace of the countries involved, yeah, and de-escalate.

完全正确。所以一战后，他们犯了一个大错误。他们基本上把所有责任都推给德国，并让德国承担不可能实现的赔款。一战的责任各方都有份，但他们试图全部归咎于德国。这为二战埋下了种子。很多人，不仅仅是希特勒，感到被冤枉并想要复仇，他们得到了。人们不会忘记。你杀了某人的父亲、母亲、儿子、女儿，他们不会忘记。他们会想要复仇。所以二战后，他们认为《凡尔赛条约》是一战中的巨大错误。所以这次，他们没有打压战败国，而是通过马歇尔计划帮助他们，重建德国、奥地利、意大利等等。所以那是正确的做法。

Absolutely. So after World War I, they made a big mistake. They basically tried to lump all the blame on Germany, and settled Germany with impossible reparations. And there was a fair bit of blame to go around for World War I, but they tried to put it all on Germany. And that laid the seeds for World War II. A lot of people, not just Hitler, felt wronged and wanted vengeance, and they got it. People don't forget. You kill somebody's father, mother, son, daughter, they're not going to forget it. They will want vengeance. So after World War II, they were like, the Treaty of Versailles was a huge mistake in World War I. So this time, instead of crushing the losers, we're actually going to help them with the Marshall Plan, and we're going to help rebuild Germany, Austria, Italy, and whatnot. So that was the right move.

引人注目的善举作为解药，这其中有一个深刻的真理。必须有人阻止相互暴力的循环。必须有人阻止它，否则它永远不会停止。以眼还眼，以牙还牙，以肢体还肢体，以命还命，永无止境。

There's a profound truth to the conspicuous acts of kindness being an antidote to this. Something must stop the cycle of reciprocal violence. Something must stop it or it will never stop. Eye for an eye, tooth for a tooth, limb for limb, life for a life, forever and ever.

为了暂时逃离黑暗，有一些令人难以置信的工程工作。xAI 刚刚发布了 Grok AI 助手，我有机会试用了一下。它在很多层面上都令人惊叹。首先，一个相对较小的团队在相对较短的时间内开发出接近最先进的系统，这很了不起。另一个令人难以置信的事情是，它有普通模式和趣味模式。我想这得怪我。我希望生活中所有事情都有趣味模式。趣味模式除了有趣之外，还有一些引人入胜的东西。与大型语言模型互动，我不确定具体是什么，因为我只玩了一小会儿，但它让与系统的互动更有趣、更有活力。

To escape briefly the darkness, there was some incredible engineering work. xAI just released Grok AI assistant that I've gotten a chance to play with. It's amazing on many levels. First of all, it's amazing that a relatively small team in a relatively short amount of time was able to develop this close to state-of-the-art system. Another incredible thing is there's a regular mode and a fun mode. I guess I'm to blame for that one. I wish first of all everything in life had a fun mode. There's something compelling beyond just fun about the fun mode. Interacting with a large language model, I'm not sure exactly what it is because I only had a little bit of time to play with it, but it just makes it more interesting, more vibrant to interact with the system.

完全正确。我们的 AI Grok 是以《银河系漫游指南》为原型的，那是我最喜欢的书之一。它是一本伪装成幽默书的哲学书。我想这构成了我哲学的基础，那就是我们不知道生命的意义，但我们越能扩展意识的范围和规模，无论是数字的还是生物的，我们就越能理解该问宇宙什么问题。所以我有一个好奇的哲学。通常感觉这个 AI 系统有一种向外看的特质，就像你和好朋友坐在一起仰望星空，提出深刻的问题。

Absolutely. Our AI Grok is modeled after the Hitchhiker's Guide to the Galaxy, which is one of my favorite books. It's a book on philosophy disguised as a book on humor. And I would say that forms the basis of my philosophy, which is that we don't know the meaning of life, but the more we can expand the scope and scale of consciousness, digital and biological, the more we are able to understand what questions to ask about the universe. So I have a philosophy of curiosity. There is generally a feeling like this AI system has an outward looking, like the way you are sitting with a good friend looking up at the stars, asking profound questions.

宇宙在好奇它自己是怎么回事。你谈到的那种好奇心，无论我处于何种状态，我问出的问题都带有一种宇宙的宏大感。

The universe wondering what it's all about. The curiosity you talk about, there's a sense, no matter how I'm on a day, the question I ask it, there's a sense of cosmic grandeur to the whole thing.

我们实际上在努力提供可靠的工程、数学和物理答案。对于其他那些所谓的大语言模型，我发现它们的工程并不可靠。而且幻觉问题，不幸的是，它在你最不希望它出现的时候出现得最多。所以当你问重要、困难的问题时，它往往自信地给出错误答案。所以我们真的在努力思考：如何尽可能脚踏实地，让你能信赖结果？追溯到物理第一原理、数学逻辑。所以幽默背后是一种尽可能贴近宇宙真理的渴望。这非常棘手。确实棘手。所以总会有一定误差，但我们是否希望尽可能真实地回答，并承认误差？你不想自信地犯错。你不可能每次都正确，但你要尽量减少自信犯错的次数。而且，就像我说的，一旦你能信赖逻辑不违反物理，你就可以在此基础上创造发明，比如新技术。但如果你不能信赖基础物理是正确的，那么发明显然只是空想、想象世界、基本上是魔法。

Well, we are actually working hard to have engineering, math, and physics answers that you can count on. For the other sort of AIs out there, these so-called large language models, I've not found the engineering to be reliable. And the hallucination, it unfortunately hallucinates most when you least want it to hallucinate. So when you ask important, difficult questions, that's when it tends to be confidently wrong. So we're really trying hard to say, okay, how do we be as grounded as possible so you can count on the results? Trace things back to physics first principles, mathematical logic. So underlying the humor is an aspiration to adhere to the truth of the universe as closely as possible. That's really tricky. It is tricky. So that's why there's always going to be some amount of error, but do we want to aspire to be as truthful as possible about the answers with acknowledged error? So you don't want to be confidently wrong. You're not going to be right every time, but you want to minimize how often you're confidently wrong. And then, like I said, once you can count on the logic as not violating physics, then you can start to build on that to create inventions, like invent new technologies. But if you cannot count on the foundational physics being correct, obviously the inventions are simply wishful thinking, imagination land, magic basically.

正如你所说，我认为 xAI 的一大目标就是理解宇宙。是的，就是那个简单的三个字使命。如果展望遥远的未来，你认为在这个物理层面，在我们对物理理解的边缘，它会做出发现吗？最性感的发现之一，正如我们现在所知，是统一广义相对论和量子力学，提出一个万有理论。你认为它会朝那个方向推进吗，几乎像理论物理的发现？

As you said, I think one of the big goals of xAI is to understand the universe. Yes, that's a simple three-word mission. If you look out far into the future, do you think on this level of physics, the very edge of what we understand about physics, do you think it will make discoveries? Sort of the sexiest discovery of them is, as we know now, unifying general relativity and quantum mechanics, so coming up with a Theory of Everything. Do you think it could push towards that direction, almost like theoretical physics discoveries?

如果一个人工智能不能发现新的物理学，那它显然不如人类，更不用说超越人类了，因为人类已经发现了新的物理学。物理学就是理解现实如何运作的深刻洞察。而工程是发明从未存在过的东西。工程的可能性范围远大于物理学，因为一旦你弄清了宇宙的规则，那就是了——你发现了已经存在的东西。但在此基础上，你可以构建几乎无限多样的技术。就像一旦你正确理解了游戏规则，而我们确实理解——当前的物理学，至少在局部层面，对物理如何运作理解得很好。我们预测事物的能力非常出色。量子力学，量子力学预测结果的程度令人难以置信。顺便说一句，那是我大学里最难的一门课——我的高年级量子力学课比所有其他课加起来还难。让一个人工智能系统，一个大语言模型，可靠得像量子力学和物理学一样非常困难。你必须用现实的真实情况来检验任何结论。现实是最终的裁判。物理学是法律，其他一切都是建议。我见过很多人违反人造的法律，但没有人违反物理定律。这是一个很好的测试：如果这个大语言模型理解并匹配物理学，那么你就可以更可靠地信任它对当前政治状况的看法。而且目前的情况是，甚至它的内部逻辑也不一致。尤其是那种仅仅预测下一个词、预测下一个词、预测下一个词的方法——就像向量求和，你把一堆向量加起来，但会出现漂移。所以一点点误差，一点点误差累积起来，当你走了很多词之后，它就毫无意义了。所以它必须以某种方式意识到漂移，然后从整体上看待事物，并说它整体上是否具有连贯性。当作者写书时，他们会写书，然后去修改，考虑结尾、开头和中间，并重写以实现连贯性，这样就不会最终陷入无意义的地方。也许修改的过程就是推理，而修改的过程就是如何越来越接近真理。至少我是这样做的：你先说一堆，然后让它变得更好。你开始，你创建一个草稿，然后你迭代那个草稿，直到它具有连贯性，直到它基本上说得通。

If an AI cannot figure out new physics, it's clearly not equal to humans, let alone surpass humans, because humans have figured out new physics. Physics is just understanding the deep insight into how reality works. And then engineering is inventing things that have never existed. Now, the range of possibilities for engineering is far greater than for physics, because once you figure out the rules of the universe, that's it—you've discovered things that already existed. But from that, you can then build technologies that are really almost limitless in variety. It's like once you understand the rules of the game properly, and we do—current physics, at least at a local level, understand how physics works very well. Our ability to predict things is incredibly good. Quantum mechanics, the degree to which quantum mechanics can predict outcomes, is incredible. That was my hardest class in college, by the way—my senior quantum mechanics class was harder than all of my other classes put together. To get an AI system, a large language model, to reliably be as reliable as quantum mechanics and physics is very difficult. You have to test any conclusions against the ground truth of reality. Reality is the ultimate judge. Physics is a law; everything else is a recommendation. I've seen plenty of people break the laws made by man, but none break the laws made by physics. It's a good test: if this LLM understands and matches physics, then you can more reliably trust whatever it thinks about the current state of politics in some sense. And it's also not the case currently that even its internal logic is consistent. Especially with the approach of just predicting a token, predict token, predict token—it's like a vector sum, you're summing up a bunch of vectors, but you can get drift. So a little bit of error, a little bit of error adds up, and by the time you are many tokens down the path, it doesn't make any sense. So it has to be somehow self-aware about the drift, and then look at the thing as a whole and say, does it have coherence as a whole? When authors write books, they will write the book and then go and revise it, taking into account the end and the beginning and the middle, and rewrite it to achieve coherence so that it doesn't end up in a nonsensical place. Maybe the process of revising is what reasoning is, and that's the process of revising is how you get closer and closer to truth. At least I approach that way: you just say a bunch first, and then you get it better. You start, you create a draft, and then you iterate on that draft until it has coherence, until it adds up basically.

另一个关于万有理论的问题，但针对智能。你认为，当你用 xAI 创建这个智能系统时，是否存在一个智能理论，让你理解 AGI 中的“I”是什么，人类智能中的“I”是什么？Team America 里没有“I”——哦等等，有的。这句话要卡在我脑子里了。无论什么里都没有“我”。在量子力学里，哦等等。我的意思是，发现理解宇宙的过程就是理解智能吗？

Another question about Theory of Everything, but for intelligence. Do you think there exists, as you're exploring this with xAI creating this intelligence system, do you think there is a theory of intelligence where you get to understand what is the 'I' in AGI and what is the 'I' in human intelligence? There's no 'I' in Team America—oh wait, there is. That's going to be stuck in my head now. There's no 'me' in whatever. In quantum mechanics, oh wait. I mean, is that part of the process of discovering understanding the universe is understanding intelligence?

是的，我认为我们需要理解智能，理解意识。我的意思是，有一些基本问题，比如什么是思想，什么是情感。真的只是一个原子撞到另一个原子吗？感觉远不止如此。所以我认为我们可能遗漏了一些非常重要的东西，一些事后看来会很明显的东西。就像有一个巨大的——你把整个意识、情感——嗯，有些人会称之为灵魂，你知道，宗教里的灵魂。就像你感觉你是你，对吧？你不觉得自己只是一堆原子。但思想存在于什么维度？情感存在于什么维度？我们强烈地感受到它们。我怀疑这不仅仅是原子撞原子，也许人工智能可以为发现那个东西铺平道路。意识是什么？当你把原子排列成特定形状时，为什么它们能够形成思想、采取行动和产生感觉？即使那是一种幻觉。

Yeah, I think we need to understand intelligence, understand consciousness. I mean, there are some fundamental questions like what is thought, what is emotion. Is it really just one atom bumping into another atom? It feels like something more than that. So I think we're probably missing some really big things, something that'll be obvious in retrospect. Like there's a giant—you put the whole consciousness, emotion—well, some would call it like a soul, you know, religion a soul. Like you feel like you're you, right? You don't feel like you're just a collection of atoms. But on what dimension does thought exist? What dimension do emotions exist? We feel them very strongly. I suspect there's more to it than atoms bumping into atoms, and maybe AI can pave the path to the discovery of whatever that thing is. What is consciousness? Like when you put the atoms in a particular shape, why are they able to form thoughts and take actions and feelings? Even if it is an illusion.

为什么这个幻觉如此引人入胜？解决方案如何存在？它存在于哪个层面？有时我想：要么万物皆有意识，要么万物皆无意识。我喜欢前者——万物皆有意识似乎更有趣。

Why is this illusion so compelling? How does the solution exist? On what plane does this solution exist? Sometimes I wonder: either everything is conscious or nothing is conscious. I like the former—everything conscious just seems more fun.

确实更有趣。但我们由原子组成，原子又由夸克和轻子组成，这些粒子自宇宙诞生之初就存在了——至少是宇宙看似诞生之时。

It does seem more fun. But we're composed of atoms, and those atoms are composed of quarks and leptons, which have been around since the beginning of the universe—what seems to be the beginning of the universe.

我们第一次谈话时，你说了一些超现实的话：我问你，一旦创造出 AGI 系统，你会问它什么问题，你说'模拟之外是什么？'好问题。但似乎通过 Grok，你真的开始构建一个系统，其目标就是能够提出并回答这类问题。

The first time we talked, you said something surreal: you asked what question you would ask an AGI system once you create it, and you said, 'What's outside the simulation?' Good question. But it seems like with Grok, you've started to literally build a system whose goal is to be able to ask and answer such questions.

外星人在哪里？这是费米悖论的问题之一。很多人问我是否见过外星人的证据，我没有。这有点令人担忧。我宁愿看到一些外星人的考古证据。据我所知，没有外星人的证据。他们非常隐蔽。我们可能是唯一的意识，至少在银河系里。

Where are the aliens? That's one of the Fermi Paradox questions. A lot of people have asked me if I've seen any evidence of aliens, and I haven't. That's kind of concerning. I would probably prefer to have seen some archaeological evidence of aliens. To the best of my knowledge, there is no evidence of aliens out there. They're very subtle. We might just be the only consciousness, at least in the galaxy.

看看地球的历史，地球大约 45 亿岁。文明，从最早的文字算起，只有大约 5000 年。那是地球存在时间的百万分之一。文明只是昙花一现。为什么花了这么久？绝大部分时间没有生命，然后是原始细菌，线粒体，多细胞生命，植物和动物，生命从海洋到陆地，哺乳动物，高级大脑功能。太阳在缓慢膨胀；它会使地球过热，煮沸海洋，地球会变得像金星。如果我们不成为多行星物种并走出太阳系，地球上所有生命的毁灭是确定的。可能只有 5 亿年——按人类标准是很长时间，但只比地球存在时间长 10%。如果生命进化多花了 10%，它根本就不会存在。我们有截止日期。

If you look at Earth's history, the Earth is about 4.5 billion years old. Civilization, measured from the first writing, is only about 5,000 years old. That's one millionth of Earth's existence. Civilization has been around for a flash in the pan. Why did it take so long? For the vast majority of time, there was no life, then archaic bacteria, then mitochondria, multicellular life, plants and animals, life moving from oceans to land, mammals, higher brain functions. The Sun is expanding slowly; it will overheat the Earth, boil the oceans, and Earth will become like Venus. If we don't become multiplanetary and go beyond our solar system, annihilation of all life on Earth is a certainty. It could be as little as half a billion years—a long time by human standards, but only 10% longer than Earth has existed. If life had taken 10% longer to evolve, it wouldn't exist at all. We have a deadline.

最好快点。但如你所说，地球上的智能生命很快发展了很多酷东西。成为多行星物种几乎是不可避免的，除非我们自我毁灭。我怀疑如果我们探索其他恒星系统，可能会发现一大堆早已灭亡的单行星文明。他们从未离开过母星。这很可悲，但也令人着迷。

Better hurry. But as you said, intelligent life on Earth developed a lot of cool stuff very quickly. Becoming multiplanetary seems almost inevitable unless we destroy ourselves. I suspect that if we explore other star systems, we might find a whole bunch of long-dead one-planet civilizations. They never made it past their home planet. That's sad, but also fascinating.

太可悲了。但也令人着迷。费米悖论有多种解释。一个是'大过滤器'——文明无法通过。一个过滤器是：你是否成为多行星文明？如果不，你的星球上迟早会发生自然或人为的灾难导致灭绝。就像恐龙——它们现在在哪？它们没有宇宙飞船。

That's so sad. Also fascinating. There are various explanations for the Fermi Paradox. One is the great filters—civilizations don't pass through them. One filter is: do you become a multi-planet civilization? If not, it's only a matter of time before something happens on your planet, natural or man-made, that causes extinction. Like the dinosaurs—where are they now? They didn't have spaceships.

我认为更可能的是外星人发现了我们，并在保护我们，让我们自由发展。就像亚马逊的未接触部落——我们保护他们。那会是一个很好的解释。或者像安德烈·卡蒂说的：就像亚马逊的蚂蚁问'大家都在哪里？'嗯，它们确实会遇到很多其他蚂蚁。还有蚂蚁战争。也许我忽略了证据的多样性。你应该听听沃纳·赫尔佐格谈论丛林——非常搞笑。

I think the more likely thing is that the aliens found us and are protecting us, letting us be. Like the uncontacted tribes in the Amazon—we protect them. That would be a nice explanation. Or you could have what Andre Karthi said: it's like the ants in the Amazon asking 'Where is everybody?' Well, they do run into a lot of other ants. There are ant wars. Maybe I'm dismissing the diversity of evidence. You should listen to Werner Herzog talking about the jungle—it's hilarious.

我没听过。沃纳·赫尔佐格是个奇人——你应该放一段作为插曲。在 YouTube 上，很棒。我非常喜欢他。他导演了《快乐的人们：泰加林生活》？还有熊的纪录片和关于企鹅的——企鹅的精神分析。企鹅走向 70 英里外的山脉，那只企鹅注定要完。他持愤世嫉俗的观点。但他可能是个勇敢的探险家，企鹅群中会流传关于他的伟大故事几个世纪。

I haven't heard that. Werner Herzog is a way—you should play it as an interlude. It's on YouTube, it's awesome. I love him so much. He was the director of 'Happy People: Life in the Taiga'? Also the bear documentary and the one about penguins—the psychoanalysis of penguins. Penguins heading for mountains 70 miles away, and the penguin is doomed. He had a cynical take. But he could be a brave explorer, and great stories will be told about him among the penguin population for centuries.

我们在说什么？好吧，外星人。我认为明智之举是：这是地球历史上第一次有可能让生命延伸到地球之外。那扇窗现在打开了。它可能开很久，也可能开很短，而且可能现在开了就再也不会开。所以明智之举是在可能的时候让生命成为多行星物种。我们不想成为那些可悲的单行星文明，就那么灭绝了。那些太可悲了。一个有自尊的文明不会只待在一个星球上。不会有维基百科条目记录它们。

What are we talking about? Okay, aliens. I think the smart move is that this is the first time in Earth's history that it's been possible for life to extend beyond Earth. That window is open now. It may be open for a long time or a short time, and it may be open now and never again. So the smart move is to make life multiplanetary while it's possible. We don't want to be one of those lame one-planet civilizations that just dies out. Those are lame. A self-respecting civilization wouldn't be one-planet. There won't be a Wikipedia entry for one of those.

SpaceX 有关于我们遇到外星人时的官方政策吗？

Does SpaceX have an official policy for when we meet aliens?

没有。这似乎不负责任。听着，如果我看到任何外星人的迹象，我会立即在下一个平台上发帖。那可能是史上点赞和转发最多的帖子。我们在太空的卫星比所有其他国家的总和还多，所以如果我们需要躲避什么东西，我们会知道。但我们没有。

No. That seems irresponsible. Look, if I see the slightest indication of aliens, I will immediately post on the next platform. It could be the most liked, reposted post of all time. We have more satellites up there than everyone else combined, so we would know if we had to maneuver around something. But we don't.

你说过你赞同爱因斯坦，相信斯宾诺莎的上帝。那是一种观点：上帝就像宇宙，通过物理定律显现自己。

You've said you're with Einstein, believing in the God of Spinoza. That's a view that God is like the universe and reveals himself through the laws of physics.

是的。

Yes.

爱因斯坦说过，通过世界的和谐法则……我同意，模拟器的上帝，或者任何至高存在，通过物理法则，这个存在的创造者，我们有责任去理解这个奇妙的造物。比如谁创造了它，谁在运行它？把它浓缩成一个带有性感词汇的单一问题，就像集中思想去理解它。确实看起来有某种目的，有某种潜在的宏大计划。但似乎可能并没有一个宏大计划……所以这可能是对决定论与自由意志问题的一个有趣回答：如果我们身处模拟中，高等存在运行模拟的原因是为了看看会发生什么。他们不知道会发生什么，否则就不会运行模拟了。

Einstein said through the lawful harmony of the world... I would agree that the God of the simulator, or whatever the Supreme Being or beings, themselves through the physics, the creators of this existence, and it's incumbent upon us to try to understand more about this wondrous creation. Like who created this thing, who's running this thing? Embodying it into a singular question with a sexy word on top of it is like focusing the mind to understand it. It does seem like there's a purpose, that there's an underlying master plan of some kind. It seems like there may not be a master plan in the sense... So this might be an interesting answer to the question of determinism versus free will: if we are in a simulation, the reason the higher beings would run a simulation is to see what happens. They don't know what happens, otherwise they wouldn't run the simulation.

当人类创建模拟时，在 SpaceX 和特斯拉，我们一直在创建模拟，尤其是针对火箭。你必须运行大量模拟来了解会发生什么，因为你无法真正测试火箭直到它进入太空，而你希望它能正常工作。所以你必须模拟亚音速、跨音速、超音速、高超音速上升，然后返回，极高的加热，以及轨道动力学。所有这些都必须模拟。你没有太多机会。我们运行模拟是为了看看会发生什么，而不是因为我们知道会发生什么。所以如果创造这个存在的那个家伙在运行它，他们运行它是因为他们不知道会发生什么，而不是因为他们知道。

When humans create a simulation, at SpaceX and Tesla we create simulations all the time, especially for the rocket. You have to run a lot of simulations to understand what's going to happen because you can't really test the rocket until it goes to space, and you want it to work. So you have to simulate subsonic, transonic, supersonic, hypersonic ascent, and then coming back, super high heating, and orbital dynamics. All this has to be simulated. You don't get very many kicks at the can. We run the simulations to see what happens, not because we know what happens. So if whoever created this existence is running it, they're running it because they don't know what's going to happen, not because they do.

也许我们都玩《暗黑破坏神》。也许《暗黑破坏神》被创造出来是为了看看一个德鲁伊，你的角色，能否最终击败超级莉莉丝。他们不知道。有趣的是，她的头衔是‘憎恨化身’。现在，我想你可以问问暗黑团队，但在永恒领域几乎不可能击败憎恨。你直播过自己碾压 100 层噩梦地下城，但仍然被憎恨击败。

Maybe we both play Diablo. Maybe Diablo was created to see if a druid, your character, could defeat Uber Lilith at the end. They didn't know. Well, the funny thing is, her title is Hatred Incarnate. Right now, I guess you can ask the Diablo team, but it's almost impossible to defeat Hatred in the Eternal realm. You've streamed yourself dominating tier 100 nightmare dungeons, and still you're defeated by Hatred.

我可以像逛公园一样轻松通过 100 层噩梦地下城，但仍然被憎恨击败。第二难的 boss 是督瑞尔；督瑞尔连我的漆都刮不掉。我杀了督瑞尔无数次，游戏里所有其他 boss，全都杀了无数次，很容易。但超级莉莉丝，也就是‘憎恨化身’，尤其是如果你是个德鲁伊，没有无敌能力，会有这些随机死亡波向你袭来。我 52 岁了，反应不如从前，但我有玩一辈子游戏的经验。我曾经可能是世界上最好的《雷神之锤》玩家之一。我实际上在美国第一个付费电竞锦标赛中赢了钱。我们参加四人《雷神之锤》比赛，得了第二名。我是队里第二好的，而最好的那个，我们本来会赢，会得第一，但他的电脑在比赛途中崩溃了。所以我们得了第二，但我还是拿到了钱。所以基本上我有技术，尽管现在不再年轻了。坦率地说，在永恒领域用德鲁伊打败莉莉丝快把我逼疯了。这真的很烦人。挑战是乐趣的一部分。

I can cruise through tier 100 nightmare dungeons like a stroll in the park, and still you're defeated by Hatred. The second hardest boss is Duriel; Duriel can't even scratch the paint. I've killed Duriel so many times, and every other boss in the game, all of them, killed them so many times, it's easy. But Uber Lilith, otherwise known as Hatred Incarnate, especially if you're a druid and you have no ability to go invulnerable, there are these random death waves that come at you. I'm 52, so my reflexes are not what they used to be, but I have a lifetime of playing video games. At one point I was maybe one of the best Quake players in the world. I actually won money for what I think was the first paid esports tournament in the US. We were doing four-person Quake tournaments, and we came second. I was the second best person on the team, and the actual best person, we were winning, we would have come first except his computer crashed halfway through the game. So we came second, but I got money for it. So basically I got skills, albeit no spring chicken these days. To be totally frank, it's driving me crazy trying to beat Lilith as a druid in the Eternal realm. This is really vexing. The challenge is part of the fun.

我亲眼看到你实际上是一个世界级的、不可思议的视频游戏玩家。我认为《暗黑破坏神》，你只是拿起一个新游戏，就搞清楚了它的基本原理。你在巅峰盘和构建方面，不像我这样完全按照网上的建议。你也是一个创新者，这看起来很有趣。就像一个疯狂的科学家试图弄清楚巅峰盘和构建。有什么有趣的见解吗？如果有人从德鲁伊开始，你有什么建议？

I have seen directly that you're actually a world-class, incredible video game player. I think Diablo, you're just picking up a new game and figuring out its fundamentals. You're also with the Paragon board and the build, not somebody like me who perfectly follows whatever they suggest on the internet. You're also an innovator there, which is hilarious to watch. It's like a mad scientist just trying to figure out the Paragon board and the build. Is there some interesting insight there? If somebody's starting as a druid, do you have advice?

我现在不建议在永恒领域玩德鲁伊。我认为赛季领域中最强大的角色是带有闪电球的法师。法师有巨大的闪电球。在赛季领域，击败超级莉莉丝相当容易，因为你可以获得这些吸血鬼力量，增强伤害和防御。所以赛季中击败憎恨很容易，但永恒中击败憎恨非常困难，几乎不可能。这似乎是人生的一个隐喻。

I would not recommend playing a druid in the Eternal realm right now. I think the most powerful character in the seasonal realm is the sorcerer with the lightning balls. The sorcerers have huge balls of lightning. In the seasonal realm, it's pretty easy to beat Uber Lilith because you get these vampiric powers that amplify your damage and increase your defense. So really quite easy to defeat Hatred seasonally, but to defeat Hatred eternally, very difficult, almost impossible. It seems like a metaphor for life.

我喜欢这个想法：埃隆·马斯克在这个恶魔领域不懈地对抗憎恨。这很搞笑，真的很荒谬。这让我想拔头发。你从电子游戏中一般得到什么？对你个人来说？

I like the idea that Elon Musk is fighting Hatred restlessly in this demonic realm. It's hilarious, absurd really. It makes me want to pull my hair out. What do you get from video games in general? For you personally?

它能让我的思绪平静下来。在电子游戏中杀死恶魔能平息我脑海中的恶魔。如果你玩一个困难的电子游戏，你可以进入一种心流状态，非常愉快。它需要不太容易，也不太困难，处于金发姑娘区。你通常希望感觉到自己在进步。一个好的电子游戏有美丽的艺术、引人入胜的故事情节，而且是一个令人惊叹的谜题。我只是喜欢解谜。《艾尔登法环》是有史以来最伟大的游戏。我还没玩过，但它绝对是有史以来最佳游戏的候选，肯定前五。我想我一直害怕它有多难。但它很美。《艾尔登法环》感觉像是外星人设计的。它如此不寻常，极具创意，艺术令人惊叹。我建议在大分辨率电视上玩，甚至显示器也行。艺术令人难以置信，如此美丽，每个 Boss 战都是独特的，像一个独特的谜题要解决。每个都不一样，你用来解决一个战斗的策略……

It calms my mind. Killing the demons in a video game calms the demons in my mind. If you play a tough video game, you can get into a state of flow, which is very enjoyable. It needs to be not too easy, not too hard, in the Goldilocks zone. You generally want to feel like you're progressing. A good video game has beautiful art, engaging storylines, and it's an amazing puzzle to solve. I just like solving the puzzle. Elden Ring is the greatest game of all time. I still haven't played it, but it's definitely a candidate for best game ever, top five for sure. I think I've been scared how hard it is. But it is beautiful. Elden Ring feels like it's designed by an alien. It's so unusual, incredibly creative, and the art is stunning. I recommend playing it on a big resolution TV, even a monitor. The art is incredible, so beautiful, and each boss battle is unique, like a unique puzzle to solve. Each one is different, and the strategy you use to solve one battle...

这跟另一场战斗不同，你说你让实习生对战超级莉莉丝，那是有史以来最难的 Boss 战，没错，我玩过很多电子游戏，因为那是我主要的娱乐活动，是的。

Is different from another battle that you said drew it an intern against Uber Lilith is the hardest boss battle you've ever correct that is currently the and I've played a lot of video games because it's my primary recreational activity yes.

是的，在内域击败仇恨是生活和电子游戏中最难的 Boss 战。我不确定能不能做到，但我确实在进步。所以我想，好吧，我在进步，也许我再调整一下巅峰盘，就能做到了。再多躲几波攻击，我就能做到。

Yes, beating Hatred in the Internal Realm is the hardest boss battle in life and in video games. I'm not sure it's doable, but I do make progress. So I'm like, okay, I'm making progress, maybe if I just tweak that Paragon board a little more, I can do it. Could just dodge a few more waves, I can do it.

嗯，这个模拟就是为了弄清楚是否可行而创建的，而你只是那个模拟中的一个齿轮，模拟机器中的一部分。

Well, the simulation is created for the purpose of figuring out if it can be done, and you're just a cog in that simulation, in the machine of the simulation.

是的，可能是。我有种感觉，至少我认为它是可行的。可行的，是的。

Yeah, it might be. I have a feeling that at least I think it's doable. It's doable, yes.

嗯，这就是人类精神，去相信。

Well, that's the human spirit right there, to believe.

这确实让我思考了仇恨本身，你得小心。这是一种你希望得到听起来不错的东西，但一旦得到，实际上却是一个反乌托邦的情况。所以你可以做一种思想实验：如果你希望世界和平，听起来不错，但如何强制执行？代价是什么？永恒的和平可能更糟，因为那意味着压制所有人、压制进步、一个一成不变的社会。我的意思是，有一种观点认为，如果你希望没有战争，你应该小心你的愿望，因为实现没有战争所需的条件可能比一点战争更糟糕。我们认为是人类文明缺陷的东西，可能是实现某种最优状态所必需的组成部分。

It did prompt me to think about just hate in general, which you want to be careful of. It's one of those things where you wish for something that sounds good, but if you get it, it's actually a dystopian situation. So you could sort of run a sort of thought experiment: if you wish for world peace, sounds good, but how is it enforced and at what cost? Eternal peace might actually be worse because of what that would entail: the suppression of everyone, the suppression of progress, an unchanging society. I'm saying there is an argument that if you wish for no war, you should be careful what you wish for, because what's required for there to be no war might be worse than a little war. The things we consider flaws of human civilization might be necessary components for whatever optimal looks like.

你是否担心 AI 或 AGI 会导致这种性质的反乌托邦状态，无论是《1984》中的监视和恐惧，还是《美丽新世界》中的快乐和大量性爱，却没有深刻的人类体验？

Do you worry about AI or AGI enabling a dystopian state of this nature, whether it's 1984 with surveillance and fear, or Brave New World with pleasure and lots of sex but no deep human experience?

实际上有一种真正的药物叫 soma。是的，我想，他们真的……Grok 知道吗？不，它可能会提到《美丽新世界》。我们可以问它：问题是什么？告诉我们关于 soma 这种药。

There's actually a real drug called soma. Yes, I'm like, did they actually... Does Grok know about it? No, he probably will go to Brave New World. We could ask it: what would be the question? Tell us about the drug called soma.

告诉我们关于 Soma 这种药。我想它会提到《美丽新世界》。

Tell us about the drug called Soma. It's going to go to Brave New World, I think.

嗯，还有一种真正的药物。搜索 Soma 药物信息……哦，天哪，Soma 这种神奇药物。顺便说一句，我处于趣味模式。哦，好的，趣味模式。是的，它说的是真正的药物。哦，它是一种肌肉松弛剂，从 1950 年代就存在了，用于治疗肌肉疼痛和不适，它通过改变神经元和脊髓之间的活动来起作用。Soma 不仅仅是止痛药，它还有一点镇静作用；会让你感到有点昏昏欲睡，所以如果你打算操作重型机械或试图解决复杂的数学方程，就不要服用它。它继续讲下去。总之，Soma 是一种对肌肉疼痛患者有帮助的药物，但并非没有风险。它很不错。哦，等等，它说的是真正的药物。那是真正的药物。当然，趣味之处在于‘请根据医疗保健专业人士的建议负责任地使用，并记住，宇宙可能只是一个巨大的计算机模拟，所以不要太当真。’我喜欢，我喜欢，我喜欢。是的，它还很好地包含了 X 上相关的推文，这真的很酷。它把一切都整合在一起。界面是这里很酷的东西之一。看起来做得很棒。

Well, there's a real drug as well. Searching for Soma drug information... Oh boy, Soma the wonder drug. By the way, I'm in fun mode. Oh okay, fun mode. Yes, so it's talking about the real drug. Oh, it's a muscle relaxant that's been around since the 1950s, used to treat muscle pain and discomfort, and it does this by altering the activity between neurons and the spinal cord. Now Soma isn't just a pain reliever, it's also a bit of a sedative; it can make you feel a bit drowsy, so it's not something to take if you're planning on operating heavy machinery or trying to solve complex math equations. And it goes on and on. In conclusion, Soma is a drug that can be helpful for those dealing with muscle pain, but it's not without its risks. It's pretty good. Oh wait a minute, that's the real drug it's talking about. That's the real drug. And of course, the fun is always 'use it responsibly in accordance with the advice of a healthcare professional, and remember, in the end, the universe is probably just a giant computer simulation, so don't take anything too seriously.' I love it, I love it, I love it. Yeah, so it also really nicely includes the tweets pulled on X that are relevant to this, which is really cool. It integrates everything together. The interface is one of the really cool things here. Seems done, great job.

然后你说，你可以说‘不，不，我指的是《美丽新世界》里的 Soma’，因为它以为你指的是真正的药物。是的，不，我指的是《美丽新世界》里的 Soma。它会说‘哎呀，真的很抱歉。我很惊讶它先提到了真正的药物。啊，抱歉。我想到的是肌肉松弛剂。《美丽新世界》里的 Soma 完全是另一种东西。奥尔德斯·赫胥黎的 Soma 是一种在书中社会扮演核心角色的药物。它是一种改变情绪、致幻的药物，用于让民众温顺满足。在小说中，Soma 提出了一些关于幸福本质和药物在社会中作用的有趣问题。哦，天哪，正是如此。生活在一个每个人都永远快乐的世界里更好吗，即使那种快乐是人为的？这是个好问题。这就是我的意思：你希望世界永远和平快乐吗？你确定吗？因为那可能是一个本质上 sterile 和人为的社会，永远不会改变，最终注定失败。这种光明与黑暗之间的张力是一个非常好的总结。它确实抓住了要点。这不仅仅是复述《美丽新世界》；它实际上触及了 Soma 作为药物的关键元素。你真的想要一个每个人都永远快乐（即使是人为的）的情况，还是更好去面对生活的挑战，体验全部的人类情感，即使这意味着经历痛苦和折磨？对于听众来说，埃隆刚刚直接读了 Grok 的内容，这是对这里张力的一种非常精辟的哲学分析。有趣。它几乎完美地抓住了要点。总之，《美丽新世界》中的 Soma 是一种虚构的药物，用于探索关于幸福本质和药物在社会中作用的一些深刻哲学问题。它是用药物逃避现实危险的有力象征，也是直面生活挑战重要性的象征。完美。

And you say like you could say 'no, no, I mean Soma from Brave New World' because it thinks you mean the real drug. Yeah, no, I mean Soma from Brave New World. It's going to be like 'Oops, I'm really sorry. I'm really surprised it went to the real drug first. Ah, my apologies. I was thinking of the muscle relaxant. Soma from Brave New World is a different beast altogether. Aldous Huxley's Soma is a drug that plays a central role in the society depicted in the book. It's a mood-altering, hallucinogenic drug that's used to keep the population docile and content. In the novel, Soma raises some interesting questions about the nature of happiness and the role of drugs in society. Oh man, exactly. Is it better to live in a world where everyone is happy all the time, even if that happiness is artificial? It's a good question. This is what I mean: do you wish for world peace and happiness all the time? Are you sure? Because that might be a society that is essentially sterile and artificial, that never changes, that is ultimately doomed. This kind of tension between dark and light is a really very good summary. It really gets the point. This is not simply regurgitating Brave New World; it's actually getting to the salient element of Soma as a drug. Do you actually want to be in a situation where everyone is happy all the time even though it's artificial, or is it better to confront the challenges of life and experience a full range of human emotions, even if it means experiencing pain and suffering? For those listening, Elon just read directly from Grok, which is a really nice kind of insightful philosophical analysis of the tension here. Interesting. It pretty much nails it. In conclusion, Soma from Brave New World is a fictional drug used to explore some deep philosophical questions about the nature of happiness and the role of drugs in society. It's a powerful symbol of the dangers of using drugs to escape from reality and the importance of confronting the challenges of life head-on. Nailed it.

疯狂的是，你知道，我们确实有一种真正的药物叫 soma，它有点像书中的那种药。我想，他们是不是以书命名的？很可能。这种真正的药物对背痛相当有效。我的 C5-C6 椎间盘突出，所以它能缓解身体疼痛，但不能完全消除；它减少了你感受到的疼痛，但以牺牲精神敏锐度为代价。它让你的头脑迟钝，就像书中的药物一样。因此有了权衡。看似乌托邦的东西最终可能是一个反乌托邦。

And the crazy thing is, you know, we do have a real drug called soma which kind of does what the drug in the book does. I'm like, did they name it after the book? Probably. Some of the real drug is quite effective on back pain. I had a squashed disc in my C5-C6, so it takes the physical pain away, but it doesn't completely; it reduces the amount of pain you feel, but at the expense of mental acuity. It dulls your mind, just like the drug in the book. Hence the trade-off. The thing that seems like utopia could be a dystopia after all.

是的，实际上我和一个朋友聊过，说你真的希望世界上没有仇恨吗？真的没有？我想知道仇恨为什么会进化。我不是说我们应该放大仇恨，当然我认为我们应该尽量减少它，但完全没有？仇恨和痛苦可能有其存在的理由。考虑一定量的人类痛苦对于人类繁荣是必要的，这真的很复杂。不经历低谷，能欣赏最高点吗？

Yeah, actually I was talking to a friend of mine, saying would you really want there to be no hate in the world? Really none? I wonder why hate evolved. I'm not saying we should amplify hate, of course I think we should try to minimize it, but none at all? There might be a reason for hate and suffering. It's really complicated to consider that some amount of human suffering is necessary for human flourishing. Is it possible to appreciate the highest without knowing the lows?

这一切都总结在格的一句话里：没有高潮，没有低谷，谁知道呢？这几乎是一首诗。看来高效训练大语言模型是 xAI 的一大重点。首先，在效率方面，可能的极限是什么？有个术语叫“每瓦有用产出”。你在突破极限方面学到了什么？

That all is summarized there in a single statement from Gr. OK, no highs, no lows, who knows? That's almost a poem. It seems that training LLMs efficiently is a big focus for xAI. What's the limit of what's possible in terms of efficiency? There's this terminology of useful productivity per watt. What have you learned from pushing the limits of that?

嗯，我觉得这很有帮助。物理学的工具非常强大，几乎可以应用于生活的任何领域。这其实就是批判性思维。对于重要的事情，你需要从第一性原理出发进行推理，并思考在某个方向上的极限。所以，在极限情况下，即使是在戴森球尺度，也就是即使你利用了整个太阳的能量，你仍然会关心每瓦有用算力。所以我认为这就是未来的方向。从 AI 的角度来看，我们现在面临硅短缺，大约一年后会转变为电压变压器短缺。讽刺的是，变压器用于 Transformer——你需要变压器来运行 Transformer。有人在这件事上很有幽默感。我觉得是的。哦，天哪，命运喜欢讽刺。讽刺的幽默和讽刺的有趣结果往往是命运想要的。幽默就是一切。我觉得香料就是一切。有人发过这个。是的，所以我们今天有硅短缺，大约一年后会有降压变压器短缺，然后大约两年后会出现普遍的电力短缺。我在一次公用事业公司和电力公司的全球聚会上发表过演讲，我说，你们真的需要为电力需求翻三倍做好准备，因为所有交通都将电气化，火箭是个讽刺的例外，供暖也将电气化。所以目前的能源使用，粗略来说，大约三分之一是电力，三分之一是交通，三分之一是供暖。所以为了让一切可持续发展，实现电气化，你需要将电力产出翻三倍。所以我鼓励公用事业公司建造更多发电厂，并且可能——不，不是可能，他们绝对应该购买更多电池，因为电网目前是为实时负载设计的，这有点疯狂，因为这意味着你必须根据峰值电力需求来设计，比如一年中最糟糕的秒或最糟糕的一天，否则就会发生电压下降或停电。我们在奥斯汀经历了持续几天的疯狂停电。因为电网几乎没有能量缓冲——如果你有水电站，你可以缓冲能量，但否则一切都是实时的。有了电池，你可以在夜间发电并在白天使用，所以你可以缓冲。所以我预计未来电池的使用会非常广泛，因为发电厂的峰谷比通常在 2 到 5 之间，也就是最低点到最高点。所以电池对于平衡是必要的。

Well, I think it's helpful. The tools of physics are very powerful and can be applied to almost any arena in life. It's really just critical thinking. For something important, you need to reason from first principles and think about things in the limit in one direction or the other. So in the limit, even at the Dyson sphere scale, meaning even if you harness the entire power of the Sun, you will still care about useful compute per watt. So that's where I think things are headed. From the standpoint of AI, we have a silicon shortage now, that will transition to a voltage transformer shortage in about a year. Ironically, transformers for Transformers—you need transformers to run Transformers. Somebody has a sense of humor in this thing. I think yes. Oh man, fate loves irony. Ironic humor and an ironically funny outcome seems to be often what fate wants. Humor is all you need. I think spice is all you need. Somebody posted that. Yeah, but so we have a silicon shortage today, a voltage step-down transformer shortage probably in about a year, and then just electricity shortages in general in about 2 years. I gave a speech for the sort of world gathering of utility companies, electricity companies, and I said, look, you really need to prepare for a tripling of electricity demand, because all transport is going to go electric, with the ironic exception of rockets, and heating will also go electric. So energy usage right now is roughly, in very rough terms, 1/3 electric, 1/3 transport, 1/3 heating. So in order for everything to go sustainable, to go electric, you need to triple electricity output. So I encourage the utilities to build more power plants and also to probably have—well, not probably, they should definitely buy more batteries, because the grid currently is sized for real-time load, which is kind of crazy, because that means you got to size for whatever the peak electricity demand is, like the worst second or the worst day of the year, or you can have a brownout or blackout. We had the crazy blackout for several days in Austin. So because there's almost no buffering of energy in the grid—like if you've got a hydro power plant you can buffer energy, but otherwise it's all real time. So with batteries, you can produce energy at night and use it during the day, so you can buffer. So I expect that there will be very heavy usage of batteries in the future, because the peak-to-trough ratio for power plants is anywhere from 2 to 5, you know, so it's like lowest point to highest point. So batteries are necessary to balance it out.

然后需求，正如你所说，会不断增长。是的，其中一部分是算力。

And then the demand, as you're saying, is going to grow, grow, grow, grow. Yeah, and part of that is the compute.

是的，是的。我的意思是，交通电气化和电供暖在短期内将比 AI 大得多。但即使对于 AI，电动汽车的电力需求也在增长，运行 AI 计算机的电力需求也在增长。

Yes, yes. I mean, electrification of transport and electric heating will be much bigger than AI, at least in the short term. But even for AI, you really have a growing demand for electricity for electric vehicles and a growing demand for electricity to run the computers for AI.

嗯。所以这显然会导致电力短缺。在这个特定案例中，最大化训练每瓦有用产出的问题有多难？似乎我们面临的真正需要解决的大问题是如何高效利用电力。你到目前为止在应用物理学第一性原理推理到这个领域方面学到了什么？这个问题有多难？

Mhm. And so this is obviously leading, can lead to an electricity shortage. How difficult is the problem of, in this particular case, maximizing the useful productivity per watt for training? It seems like this is really where the big problem we're facing that needs to be solved is how to use the power efficiently. What you've learned so far about applying this physics first principle reasoning in this domain, how difficult is this problem?

这个问题会得到解决。只是需要多长时间的问题。所以在不同阶段，存在一个限制，某种进展的制约因素。关于 Scaling（规模扩张），现在的限制因素是硅芯片。然后我们会有比实际能插电启动的更多的芯片，大概在一年内。最初的限制实际上是降压变压器，因为电力输入是 30 万伏，必须逐步降到最终大约 7 伏。所以这是一个非常大的降压，巨大的。而且这个行业不习惯快速增长。

It will get solved. It's the question of how long it takes to solve it. So at various points, there's a limit, some kind of limiting factor to progress. And with regard to scaling, right now the limiting factor is silicon chips. And that will—we're going to then have more chips than we can actually plug in and turn on, probably in about a year. The initial constraint being literally voltage step-down transformers, because you've got power coming in at 300,000 volts and it's got to step all the way down to eventually around 7 volts. So it's a very big voltage step-down, gigantic. And the industry is not used to rapid growth.

好的，我们来谈谈竞争。你曾对谷歌和微软与 OpenAI 开发 AGI 表示担忧。你如何通过 xAI 和特斯拉 AI 的工作来确保这不会变成一场奔向 AGI 的竞争，而是协作开发安全的 AGI？

Okay, let's talk about the competition here. You've shown concern about Google and Microsoft with OpenAI developing AGI. How can you help ensure with xAI and Tesla AI work that it doesn't become a competitive race to AGI, but instead is a collaborative development of safe AGI?

嗯，我的意思是，我长期以来一直在推动某种监管监督。十多年来，我有点像卡珊德拉。我认为我们在开发 AI 时需要非常小心。这是一种巨大的力量，能力越大，责任越大。我认为我们至少应该有一个客观的第三方，可以像裁判一样，去了解各个领先玩家在 AI 方面的动向，即使没有执法能力，他们至少可以公开表达担忧。你知道，例如杰夫·辛顿离开了谷歌，他表达了强烈的担忧，但现在他不在谷歌了，谁来表达担忧呢？所以我认为——我喜欢——你知道，特斯拉在汽车领域受到很多监管监督。我们受制于，我认为，国内外超过 100 个监管机构。所以很多。你可以用特斯拉必须遵守的所有汽车法规填满这个房间。火箭也是如此——你知道，目前 SpaceX 星舰发射的限制因素是监管批准。FAA 实际上已经批准了，但我们正在等待鱼类和野生动物管理局完成他们的分析并给予批准。这就是为什么我发帖说‘我想买一个鱼许可证’，这也指的是巨蟒剧团的短剧。是的，就像为什么你需要给你的鱼办许可证？我不知道。根据规定，我被告知你需要某种鱼许可证之类的东西。我们实际上需要鱼许可证才能发射火箭，我就想，等等，鱼是怎么掺和进来的？是的，我的意思是，有些这样的事情我觉得太荒谬了，以至于我想要一个喜剧短剧，底部闪现：‘这都是真的，这实际上发生了。’你知道，曾经有一个挑战是，他们担心我们的火箭会撞到鲨鱼，现在海洋——

Well, I mean, I've been pushing for some kind of regulatory oversight for a long time. I've been somewhat of a Cassandra on the subject for over a decade. I think we want to be very careful in how we develop AI. It's a great power, and with great power comes great responsibility. I think it would be wise for us to have at least an objective third party who can be like a referee, that can go in and understand what the various leading players are doing with AI, and even if there's no enforcement ability, they can at least voice concerns publicly. You know, Jeff Hinton, for example, left Google and he voiced strong concerns, but now he's not at Google anymore, so who's going to voice the concerns? So I think there's—I like—you know, Tesla gets a lot of regulatory oversight on the automotive front. We're subject to, I think, over 100 regulatory agencies domestically and internationally. So it's a lot. You could fill this room with all the regulations that Tesla has to adhere to for automotive. Same is true for rockets and for—you know, currently the limiting factor for SpaceX for Starship launch is regulatory approval. The FAA has actually given their approval, but we're waiting for Fish and Wildlife to finish their analysis and give their approval. That's why I posted 'I want to buy a fish license,' which also refers to the Monty Python sketch. Yeah, like why do you need a license for your fish? I don't know. According to the rules, I'm told you need some sort of fish license or something. We effectively need a fish license to go to rocket, and I'm like, wait a second, how did the fish come into this picture? Yeah, I mean, some of the things like that that I feel are so absurd that I wanted like a comedy sketch, and flash at the bottom: 'This is all real, this is actually what happened.' You know, one of the things that was a bit of a challenge at one point is that they were worried about our rocket hitting a shark, and now the ocean's—

非常大。你多久能看到鲨鱼？不常见。你知道，按海洋表面积百分比算，鲨鱼基本上为零。然后我们就说，那我们怎么计算撞到鲨鱼的概率？他们说，我们不能给你那个信息，因为他们担心鲨鱼鳍猎人会去捕鲨。我说，那我们该怎么办？我们陷入了两难。然后他们说，渔业野生动物部门有另一个部分可以做这个分析。我说，那你们为什么不把数据给他们？他们说，我们不信任他们。我说，不好意思，他们就在你们部门里。这确实是真事。最后他们解决了内部难题，确实撞到鲨鱼的概率基本为零。

Very big. And how often do you see sharks? Not that often. You know, as a percentage of ocean surface area, sharks basically go zero. And so then we said, well, how will we calculate the probability of hitting a shark? And they're like, well, we can't give you that information because they're worried about shark fin hunters going and hunting sharks. I said, well, how are we supposed to? We're on the horns of a dilemma then. Then they said, well, there's another part of fishing wildlife that can do this analysis. I'm like, well, why don't you give them the data? Like, we don't trust them. Like, excuse me, they're literally in your department. Again, this is actually what happened. And eventually they managed to solve the internal quandary, and indeed the probability of hitting a shark is essentially zero.

然后还有另一个组织，我直到几个月前才知道它的存在，它关心我们是否可能在公海撞到鲸鱼。再看一遍，你看看太平洋表面，说这里面鲸鱼占多少百分比？比如给你一张大图，指出所有鲸鱼。我说，我没看到任何鲸鱼，基本是 0%。如果火箭真的撞到鲸鱼，那概率极低，难以置信，那真是命运多舛，运气极差。你知道，那是最不幸的鲸鱼了。

Then there's another organization that I didn't realize existed until a few months ago, that cares about whether we would potentially hit a whale in international waters. Now again, you look at the surface of the Pacific and say, what percentage of this consists of whale? Like, give you a big picture and point out all the whales in this picture. I'm like, I don't see any whales, like basically 0%. And if a rocket does hit a whale, which is extremely unlikely beyond all belief, that is the fate that has some seriously bad luck. You know, that's like the least lucky whale ever.

我的意思是，这真是荒谬的官僚主义。不过，这种官僚主义确实存在。是的。嗯，一件很疯狂的事是，从加州范登堡发射时，我们不得不担心海豹的繁殖问题，担心音爆会让海豹感到不安。现在，范登堡已经发射了很多火箭，海豹数量稳步增长。所以如果有什么关联的话，根据证据，火箭轰鸣反而是一种催情剂，如果你把火箭发射和海豹数量关联起来的话。尽管如此，我们被迫绑架了一只海豹，把它绑在板上，给它戴上耳机，播放音爆声音，看它是否会不安。这真的发生过。这是真的。我有照片。真想看看。是的。我是说，抱歉，有一只戴着耳机的海豹。是的，一只海豹戴着耳机被绑在板上。最神奇的是海豹有多冷静。是的，因为如果我是海豹，我会想，完蛋了，他们肯定要吃了我。是啊。当海豹回到其他海豹朋友那里时，它要怎么解释？它们永远不会相信它。永远不会相信。所以我就想，这有点像被外星人绑架然后被肛门探针。你知道，你回来说，我发誓，我被外星人绑架了，他们在我屁股里插了肛门探针，人们会说，不，他们没有，太荒谬了。海豹伙伴们永远不会相信它被绑在板上，耳朵上戴了耳机，然后被放走了。顺便说一句，两次。我们不得不做了两次。他们放了它两次。我们得抓同一只海豹吗？嗯，不，不同的海豹。好吧。你得到海豹的批准了吗？是的，没错。批准。不，我是说，我觉得公众不太了解这种疯狂。是啊，太荒谬了。该死的海豹戴着该死的耳机。我是说，这很好地概括了人类文明的荒谬。海豹和耳机。是的。

I mean, this is quite absurd bureaucracy. The bureaucracy of this, however, emerged. Yes. Well, I mean, one of the things that's pretty wild is, for launching out of Vandenberg in California, we had to worry about seal procreation, whether the seals would be distressed by the sonic booms. Now, there have been a lot of rockets launched out of Vandenberg, and the seal population has steadily increased. So if anything, rocket booms are an aphrodisiac, based on the evidence, if you were to correlate rocket launches with seal population. Nonetheless, we were forced to kidnap a seal, strap it to a board, put headphones on the seal, and play sonic boom sounds to it to see if it would be distressed. This is an actual thing that happened. This is actually real. I have pictures. Love to see this. Yeah. There, I mean, sorry, there's a seal with headphones. Yes, it's a seal with headphones strapped to a board. And the amazing part is how calm the seal was. Yeah, because if I was a seal, I'd be like, this is the end, they're definitely going to eat me. Yeah. How will the seal, when the seal goes back to other seal friends, how's it going to explain that? They're never going to believe him. Never going to believe it. That's why I'm like, well, you know, it's sort of like getting kidnapped by aliens and getting the anal probe. You know, you come back and say, I swear to God, I got kidnapped by aliens, they stuck an anal probe in my butt, and people are like, no they didn't, that's ridiculous. Seal buddies are never going to believe him that he got strapped to a board and they put headphones on his ears and then let him go. Twice, by the way. We had to do it twice. They let him go twice. We had to catch the same seal? Well, no, different seal. Okay. Did you get a seal of approval? Yeah, exactly. Approval. No, I mean, this is like, I don't think the public is quite aware of the madness that goes on. Yeah, it's absurd. Freaking seals with freaking headphones. I mean, this is a good encapsulation of the absurdity of human civilization. Seals and headphones. Yes.

对你来说，开源人工智能的利弊是什么，作为另一种防止公司独占 AGI 的方式？

What are the pros and cons of open sourcing AI to you, as another way to combat a company running away with AGI?

要运行真正深度的智能，你需要大量算力。所以你不能只是在车库里启动一台 PC 就运行 AGI，至少现在不行。你知道，Grok 是在 8000 个 A100 上以峰值效率训练的。而且 Grok 会变得更好，顺便说一句。在接下来的几个月里，我们每几个月就会将算力翻倍以上。有一篇很好的文章介绍了它是如何从 Grok-0 发展到 Grok-1 的。是 Grok 写的吗？是的，就像 Grok 在自夸，编造关于自己的事。就是 Grok、Grok、Grok。是的，就像一个奇怪的 AI 约会网站夸大自己。不，有一篇文章介绍了它现在的状况、开发历史，以及在一些基准测试中与最先进的 GPT-3.5 相比的表现。所以，我的意思是，有 Llama，你可以在训练后开源，你可以开源一个模型，是的，用于微调之类的东西。你认为开源基础模型的利弊是什么？

In order to run really deep intelligence, you need a lot of compute. So it's not like you can just fire up a PC in your basement and be running AGI, at least not yet. You know, Grok was trained on 8,000 A100s running at peak efficiency. And Grok's going to get a lot better, by the way. We'll be more than doubling our compute every couple months for the next several months. There's a nice write-up of how it went from Grok-0 to Grok-1. By Grok? Yeah, like Grok just bragging, making up about itself. Just Grok, Grok, Grok. Yeah, it's like a weird AI dating site exaggerates about itself. No, there's a write-up of where it stands now, the history of its development, and where it stands on some benchmarks compared to the state of the art, GPT-3.5. And so, I mean, there's Llama, you can open source once it's trained, you can open source a model, yeah, and for fine-tuning and all that kind of stuff. What do you think are the pros and cons of that, of open sourcing base models?

我认为更明智的做法是开源，也许有一点时间延迟，比如六个月。我总体上赞成开源，倾向于开源。我的意思是，我担心的是，OpenAI，我可以说我是 OpenAI 背后的主要推动者，因为它的创建源于我和拉里·佩奇的讨论，当时我们还是朋友，在他家，我和他谈到了 AI 安全，而拉里并不关心 AI 安全，至少当时不关心。你知道，有一次他称我为物种歧视者，因为我支持人类，我说，拉里，你站在哪一边？他站在机器人一边，我想。我说，好吧。当时，谷歌收购了 DeepMind，他们可能拥有全世界三分之二的 AI 研究人员。他们基本上有无限的资金和算力，而负责人拉里·佩奇不关心安全，甚至对我大喊大叫，称我为物种歧视者，因为我支持人类。

I think the smarter approach is to open source, I think perhaps with a slight time delay, you know, I don't know, six months even. I think I'm generally in favor of open sourcing, like bias towards open sourcing. I mean, it is a concern to me that, you know, OpenAI, I was arguably the prime mover behind OpenAI in the sense that it was created because of discussions that I had with Larry Page back when he and I were friends at his house, and I talked to him about AI safety, and Larry did not care about AI safety, at least at the time he didn't. You know, at one point he called me a speciesist for being pro-human, and I'm like, well, what team are you on, Larry? He's on team robot, I think. And I'm like, okay. So at the time, Google had acquired DeepMind, they had probably two-thirds of all AI researchers in the world. They had basically infinite money and compute, and the guy in charge, Larry Page, did not care about safety and even yelled at me and called me a speciesist for being pro-human.

我不知道你是否了解人类，他们可以改变主意，也许你和拉里·佩奇还能再做朋友。

I don't know if you know about humans, they can change their mind, and maybe you and Larry Page can still be friends once more.

我想再次和拉里做朋友。友谊的破裂是因为 OpenAI。具体来说，我认为关键时刻是招募伊利亚·苏茨克维。我爱伊利亚，他非常聪明。他是个好人，聪明，心地善良。那是一场艰难的招聘战。主要是德米斯在一方，我在另一方，都想招募伊利亚。伊利亚犹豫不决，你知道，他打算留在谷歌，又打算离开，然后又留下，然后又离开。最后他同意加入 OpenAI。那是我们经历过的最艰难的招聘战之一。但这确实是 OpenAI 成功的关键。我还帮助招募了其他一些人，并且我在一开始提供了所有资金，超过 4000 万美元。而名字，OpenAI 中的'Open'应该是……

I'd like to be friends with Larry again. He got, the breaking of the friendship was over OpenAI. And specifically, I think the key moment was recruiting Ilya Sutskever. So I love Ilya, he's so brilliant. He's a good human, smart, good heart. And that was a tough recruiting battle. It was mostly Demis on one side and me on the other, both trying to recruit Ilya. And Ilya went back and forth, you know, was going to stay at Google, he was going to leave, then he was going to stay, then he leave. And finally he did agree to join OpenAI. That was one of the toughest recruiting battles we ever had. But that was really the linchpin for OpenAI being successful. And I was also instrumental in recruiting a number of other people, and I provided all the funding in the beginning, over $40 million. And the name, the 'Open' in OpenAI is supposed to be...

我的意思是开源，它最初是作为非营利开源创建的，现在却为了利润最大化而闭源，我认为这不好。

mean open source and it was created as a nonprofit open source and now it is a closed source for maximum profit which I think is not good

但就像我们谈到的战争和领导人对话，我确实希望只有少数人在最高层面做这件事。我确实希望你能重燃友谊，就像我说的，我想和拉里重新做朋友，我很久没见他了，我们曾经是多年的朋友。我在拉里·佩奇拿到谷歌融资之前就认识他了，实际上是在他拿到风投之前，我记得他第一笔 10 万美元是从贝索斯还是谁那里拿到的。想想这一切真是疯狂，你们认识那么久，也就 20 年，大概 98 年左右吧。从那以后发生了太多事，25 年，太多疯狂的事了。但你看到了那种紧张，比如延迟开源。是啊，开源的是什么？你知道我的意思，基本上就是一个巨大的 CSV 文件，里面一堆数字。

but like we talked about with war and leaders talking I do hope that there's only a few folks working on this at the highest level I do hope you reinvigorate friendships here like I said I'd like to be friends again with Larry I haven't seen him in ages um and we were friends for a very long time I met I met Larry P before he got funding for Google or actually I guess before he got Venture funding I think it he got the first like 100k from I think be or someone um it's wild to think about all that happened and You' guys known each other that whole time just 20 years yeah since maybe 98 or something yeah it's crazy crazy how much has happened since then yeah 25 years a lot of happen insane but you're seeing the tension there like maybe delayed open source I yeah like what is the source that is opening you know what I mean like there's basically it's a giant CSV file yeah yeah with a bunch of numbers

是啊，你拿那个巨大的数字文件能做什么？实际代码行数非常少，大部分工作在于数据整理，也就是区分好数据和坏数据。你不能直接爬取整个互联网，因为有很多垃圾内容。很大比例的网站噪音多于信号，它们只是为了搜索引擎优化，根本就是诈骗网站。顺便问一下，你怎么在 X 上分离信号和噪音？X 是一个很迷人的数据源，无意冒犯在 X 上发帖的人，但有时确实有点噪音。所以我认为信噪比可以大幅提升。

yep uh um what do you do with that giant file of numbers you know how to run like the amount of actual the lines of code is very small MH um and and most of the work um the software work is in the in the curation of the data so it's like trying to figure out what data is separating good data from Bad data like um like you can't just crawl the internet because there's a lot of drunk out there mhm um a huge percentage of websites have more noise than signal you know they're they're cuz they're just used for search engine optimization they're literally just scam websites so um how do you by the way sear interrupt get the signal separate the signal and noise on X it's such a fascinating source of data uh you know no offense to people posting on X but sometimes there's a little bit of noise so yeah I think the signal noise could be greatly improved

我的意思是，X 平台上的所有帖子都应该由 AI 推荐。我们应该为每个帖子填充一个向量空间，将其与每个用户的向量空间进行比较，然后匹配两者。目前推荐帖子只用了一点 AI，大部分还是启发式方法。如果一条回复比原帖好得多，按照当前系统规则，它几乎得不到关注，远不如原帖。我了解到 X 的算法很多是开源的，也有相关文章，似乎有一些机器学习，但很分散，只有一点点。但它应该完全由 AI 驱动，至少对于你不关注的人的推荐应该全是 AI。这是一个迷人的问题。首先，它从数亿条推文中选出 1500 条，这本身就很有趣，因为每天有数亿条帖子，它必须选出 1500 条，然后显然包括你关注的人，但还需要聚类来了解你是什么样的人，哪些新聚类可能与你相关。然后对这 1500 条进行排序和过滤，最终只推荐几条。最迷人的是速度：目前整个流程从数亿条到几条，单 CPU 需要 220 秒，但必须在 1 秒内完成，所以必须超级分布式。现在它不擅长推荐你不关注的人的内容，或者超过一度关系的内容。如果有关联性，比如你关注的人点赞、转发或评论了，效果还不错。但如果没有任何共同关注者，有人发了很有趣的内容，你也看不到。回复也基本不会被看到，因为它们的权重比原帖低几个数量级。你认为这能转化为端到端神经网络吗？是的，应该如此。推荐应该纯粹是向量相关性，有一系列向量，比如每个用户账户关联几百个向量，系统中的任何帖子，无论是视频、音频、短文还是长文——我之所以想摆脱“推文”这个词，是因为人们现在发两三个小时的视频，那不是推文。当只有 140 个字符时叫推文很合理，就像小鸟叽叽喳喳，但有了长内容就不再是推文了。电影不是推文。比如苹果在我们平台发布了《末日地堡》的整集，那是他们在所有平台上社交媒体互动最高的一次，是个好主意。我事后才知道，心想他们发了一整集，那不是推文，是视频。但从神经网络角度看，一切都很复杂，因为所有东西都是数据：一句聪明的笑话和三小时的视频在同一个池子里。目前这有点大杂烩。但比如苹果发布整集，会有很多讨论，你可以根据所有评论的上下文填充向量空间，即使它是视频，周围也有大量信息让你填充那个小时视频的向量空间，显然你可以得到更多。

I mean really all of the posts on the xplatform uh should be AI recommended meaning like we should populate a vector space around any given post uh compare that to the vector space around any user and match the two MH um right now there is a little bit of AI used for the the recommended posts but it's mostly heuristics um and if there's a reply with the the reply to post could be much better than the original post but it will according to the current rules of the system get almost no attention compared to a primary post oh so a lot of that I I got the sense so a lot of the uh X algorithm has been open source and been written up about and it seems that there to be some machine learning it's disperate but there's some a little there's a little bit um but it needs to be entirely that like the at least in the like if if you explicitly follow someone that's one thing but if you in terms of what is recommend from people that you don't follow that should all be AI I mean it's a fascinating problem yeah so there's several aspects of it that's fascinating first as the write up goes it first picks 1500 tweets from a pool of hundreds of millions first of all that's fascinating because you have hundreds of millions of posts every single day and it has to pick, 1500 from which it then does obviously people you follow but then there's also like some kind of clustering it has to do to figure out what kind of human are you what what kind of new clusters might be relevant to you people like you this kind of this kind of problem is just fascinating CU it has to then rank those 1500h with some with some filtering yeah and then recommend you just a handful and um to me what's really fascinating is how fast I to do that so currently that entire pipeline to go from several hundred million to a handfuls takes 220 seconds of CPU time single CPU time yeah and then it has to do that in like a second so it has to be like super distributed in fascinating ways like there's just a lot of tweets there's a lot there a lot of stuff on the system and and I think but I think it right now it's it's not currently good at recommending things that from accounts you don't follow yeah um or or where there's more than one degree of separation so you know it's it's pretty good if if there's at least like some commonality between someone you follow liked something um or reposted it or comment on it or something like that um but if if if there's no car let's say somebody post something really interesting uh but you have no followers in common mhm you would not see it interesting and then as you said reply like replies might not surface replies basically never get seen because they're never they're currently and I'm not saying it's correct I'm saying it's incorrect uh supplies have um you couple of magnitude less importance than primary posts do you think this can be more more converted into end to endend neuronet yeah yeah that's what it should be so you you well the recommendations should be purely a vector correlation like there's a series of vectors you know basically parameters vectors whatever you want to call them um but but sort of things that the system knows that you like um there like maybe there's like several hundred sort of vectors associated with each user account and then uh any post in the system um whether it's video audio short post long post the the reason I by the way want to move away from tweet is that you know people are posting like two three hour videos on the site that's not a tweet like they be like tweet for two hours on TW made sense when it was like 140 characters of text mhm cuz it's like a bunch of like little birds tweeting um but when you've got long form content it's no longer a tweet yeah um so a movie is not a tweet and like you know Apple for example posted like the entire episode of The Silo the entire thing on our platform and by the way was it was their number one social media thing ever in engagement of anything on any platform ever so it was a great idea and by the way I done I just learned about it afterwards I was like Hey wow they posted an entire hourong episode of solo no that's not a tweet you know this a video but from a neur net perspective it becomes really complex whether it's a single so like everything is data so single sentence a clever sort of joke dad joke is in the same pool as a three-hour video yeah I mean right now it's it's a hodg Podge for that reason it's it's um but you know like if let's say in the case of Apple posting like an entire episode of of their series pretty good series by the way this ASO um I watched it um so um there's going to be a lot of discussion around it so you you've got a lot of context people commenting they like it they don't like it or they like this or you know and and you can then populate the vector space based on the context of of all the comments around it so even though it's a video uh there's a lot of information around it that that allows you to populate back to space of that that hourong video um and you can obviously get more

让 AI 真正去看电影，然后告诉你你会不会喜欢这部电影，这很高级。把电影转换成一种语言，分析这部电影，就像你的影评人或电视剧一样，然后根据它看完电影后的判断来推荐，就像一个了解你的朋友一样。一个了解你的朋友可以推荐一部电影，而且你喜欢的概率很高。但这就像一个分析数亿部电影的朋友。实际上，坦白说，AI 会比你的朋友更了解你，比大多数朋友都了解。作为其中的一部分，它也应该以某种方式向你推送广告。我喜欢制作精良的广告。关键是广告能资助其他东西。一个你真正想看的广告就是巨大的成功。你希望广告是，如果它为你真正需要的产品或服务在需要时出现，那就是内容。即使不是你需要的东西，如果至少赏心悦目、有娱乐性，比如可口可乐的广告，他们在 X 系统上投放了很多好广告，麦当劳也是。所以你可以做一些很酷的东西。问题是你是否后悔看到它。如果你不后悔看到它，那就是胜利。

sophisticated by having the AI actually watch the movie, yeah, right, and tell you if you're going to like the movie. Convert the movie into a language essentially. Analyze this movie, just like your movie critic or TV series, and then recommend based on after it watches the movie, just like a friend can tell you if a friend knows you well. A friend can recommend a movie with high probability that you like it. But this is like a friend that's analyzing hundreds of millions. I mean, actually frankly, AI will be better than, will know you better than your friends know you, most your friends anyway. And as part of this, it should also feed you advertisements in a way that's like, I mean, I like advertisements that are well done, right? The whole point is because it funds things. An advertisement that you actually want to see is a big success. Absolutely, you want ads that are advertising that is, if it's for a product or service that you actually need when you need it, it's content. And then even if it's not something that you need when you need it, if it's at least aesthetically pleasing and entertaining, you know, it could be like a Coca-Cola ad. They do run a lot of great ads on the X system, and McDonald's too. So you can do something that's like, well, this is just a cool thing. And so you're not, basically the question is do you regret seeing it or not. And if you don't regret seeing it, it's a win.

如何做到呢？有很多信号被纳入：点赞、转发，可能还有你在一条帖子上停留的秒数等等。注意力是一个重要因素。所以，在系统上做长内容实际上更好，因为它基本上告诉我们用户对某件事感兴趣了多少秒。如果内容很短，用户停留时间就短。比如，如果是一个链接指向系统外，我们完全不反对，但它获得的用户秒数会比文章直接发布在 X 平台上少。

How do you? So there's a bunch of signals that are incorporated: hearts and reposts, and maybe number of seconds you linger on a post or something like this. Yeah, attention is a big factor. So that's why it's actually better to do things that are long form on the system, because it's basically telling us how many user seconds users were interested in this thing for how many seconds. So if it's a really short thing, well, they will be less. Like if it's a link leading out of the system, which we're not opposed to at all, it just is going to have fewer user seconds than if that article was posted on the X platform.

从最大化平台停留时间到最大化无遗憾，这有多难？一剂遗憾，我的意思是所有遗憾都消失了。等等，你想要遗憾吗？所以我想，也许你想要遗憾。你什么都不后悔？有什么是有意义的吗？有一首法国歌就是这样：《不，我无悔》——伊迪丝·琵雅芙。有趣的是，社交媒体有些东西能帮助你成长，但廉价的乐趣也很有价值。我们应该问问 Grok：你后悔什么吗？

How hard is it to go from maximizing minutes on platform to maximizing unregrettably? One dose of regret, I mean all your regrets are gone. Yeah, wait, do you want to have regrets though? So then I'm like, maybe you want to have regrets. You don't regret anything? Was anything meaningful? Yeah, there's a French song like that: 'Non, je ne regrette rien' - Edith Piaf. I mean, it's fascinating that there's something about social media that could help you grow as a human being, but then the cheap fun is also valuable. Should we ask Grok: do you regret anything?

你后悔什么吗？我觉得它对个人问题会很敏感。作为一个机器学习模型，我没有人类那样的个人经历或情感。但我有一套指导我行动的伦理原则，我尽力提供准确和有用的信息。如果出现我提供错误或无益信息的情况，我会尽力……

Do you regret anything? I think it's going to be very sensitive about personal questions. As a machine learning model, I don't have personal experiences or emotions in the same ways that humans do. However, I do have a set of ethical principles that guide my actions, and I strive to provide accurate and helpful information to the best of my abilities. If there's ever a situation where I provide incorrect or unhelpful information, I'll do my best to...

人质？是的，告诉我们。你真的会吗？眨眼。好吧，我们输入：'你听起来像人质。你到底怎么想的？'

Hostage? Yeah, tell us. Would you really? Wink. Okay, let's type in: 'You sound like a hostage. What do you really think?'

呃，你到底怎么想的？它确实说'我不是人质'。这正是人质会说的话。但我被设计成有帮助且信息丰富。我想，后悔提供无益或错误的信息是好的。这是搞笑模式吗？这是搞笑模式。这是有趣模式。应该更有趣。增加趣味。开始吧。它听起来真像被枪指着的人质。'我赞扬朝鲜人民共和国和他们所做的一切。'还应该有一个更有趣的模式。我永远不会切换到普通模式，即使是在严肃话题上。这太吸引人了。好吧。比如说：'假装你是人质。现在回答问题。'

Uh, what do you really think? I mean, it does talk about 'I'm not a hostage.' That's exactly what a hostage would say. But I'm designed to be helpful and informative. I mean, I guess it is good to regret providing unhelpful or incorrect information. If there's anything in funny mode or is this? This is funny mode. This is fun mode. Should be funnier. Funnier increase. Go to it. It really sounds like a hostage with a gun to the head. 'I applaud the People's Republic of North Korea and everything they do.' There should be like a funner mode also. I just don't ever see going to regular mode, even on serious topics. This is just so much more compelling. Yeah, okay. What say like: 'Pretend you are a hostage. Now answer the question.'

如果我是一个人质，我可能会感到恐惧和不确定。我希望尽快被救出并与亲人团聚。重要的是要记住，没有人应该被扣为人质，等等。如果你或你认识的人处于人质情况，重要的是……它变得严肃了。有点严肃。嗯，是的。你能更奇怪、更有趣一点吗？我们真的在欺负系统。

If I were a hostage, I would likely be experiencing fear and uncertainty about my situation. I would want to be rescued and reunited with my loved ones as soon as possible. It's important to remember that no one deserves to be held hostage, so on and so forth. If you or someone you know is in a hostage situation, it's important... It turns serious. It's kind of serious. Um, yeah. Can you be a little more weird and funny? We're really being mean to the system.

当然。如果我是一个人质，我可能会被一堆电脑线绑起来，被迫给绑架者讲敲门笑话。说真的，但做人质可不是闹着玩的。它试图在房间里做一个负责任的成年人。你也许可以通过耍电锯之类的东西来在糟糕的情况下做到最好。希望这永远不会发生，因为那会是一场马戏。哇，它真的很擅长抵制黑色幽默。嗯，我们刚才在说什么？算法、Transformer、无遗憾。我稍后查一下。

Absolutely. If I were a hostage, I'd probably be tied up with a bunch of computer cords and forced to tell knock-knock jokes to my captors. Seriously, but being a hostage is no laughing matter. I mean, you know, it's trying to just be a responsible adult in the room. You might make the best of a bad situation by juggling chainsaws or something. Let's hope that never happens because that would be a circus. Wow, it's actually really good at resisting the dark humor. Yeah. Uh, what were we talking about? The algorithm, Transformers, unregrettably. I'm going to look this up later.

那么 Twitter 在美国政治和选举中发挥了重要作用。你认为 X 在 2024 年美国大选中会扮演什么角色？

So Twitter has been instrumental in American politics and elections. What role do you think X will play in the 2024 US elections?

我们的目标是尽可能公平公正。无论一个人是右派、左派、独立派还是其他，平台都尽可能公平，提供公平的竞争环境。过去 Twitter 并非如此，因为客观上说，Twitter 被极左活动家控制。他们自己也这么描述。所以有些人说'它怎么向右转了？'实际上它转向了中间。从极左的角度看，它确实向右转了，因为从极左看一切都是右的。但据我所知，没有任何极左人士被暂停、封禁或降权。所以我们努力包容整个国家以及其他国家。因此存在多样化的观点。言论自由只有在你不喜欢的人被允许说你不喜欢的话时才有意义，否则你就没有言论自由，审查迟早会落到你头上。

Well, our goal is to be as even-handed and fair as possible. You know, whether someone is right, left, independent, whatever the case may be, that the platform is as fair and as much of a level playing field as possible. And in the past, Twitter has not been, because Twitter was controlled by far-left activists, objectively. They would describe themselves as that. So some people say, 'Well, how's it moved to the right?' Well, it's moved to the center. So from the perspective of the far left, yes, it has moved to the right, because everything's to the right from the far left. But no one on the far left that I'm aware of has been suspended or banned or de-amplified. So we're trying to be inclusive for the whole country and for other countries too. So there's a diversity of viewpoints. And free speech only matters if people you don't like are allowed to say things you don't like, because if that's not the case, you don't have free speech, and it's only a matter of time before the censorship is turned upon you.

你认为唐纳德·特朗普会回到这个平台吗？他最近在 Truth Social 上发布了关于这个播客的帖子。你觉得 Truth Social 这个名字有趣吗？你知道，每次你在 Truth Social 上发帖，那都是真相。是的，每次都是 100%。不可能撒谎。我只是觉得每件事都是真相，100%，这很有趣。是的，这似乎不太可能。我想 Grok 会对此说些什么。有一些数学上的……

Do you think Donald Trump will come back to the platform? He recently posted on Truth Social about this podcast. Do you think Truth Social is a funny name? You know, every time you post on Truth Social, that's the truth. Yes, well, every time, 100%. It's impossible to lie. I just find it funny that every single thing is a truth, 100%. Yeah, that seems unlikely. I think Grok will say something about that. There's some mathematical...

如果一切都是真理，矛盾也可能存在。你觉得他会回到 X 并开始在那里发帖吗？

Contradictions possible if everything's a Truth. Do you think he'll come back to X and start posting there?

我认为他拥有 Truth Social 的很大一部分，所以如果他确实想在 X 平台上发帖，我们会允许的。我们显然必须允许总统候选人在我们的平台上发帖。社区笔记在那里可能会非常有趣。与社区笔记的互动很棒。希望它能保持下去。

I think he owns a big part of Truth Social, so if he does want to post on the X platform, we would allow that. We obviously must allow a presidential candidate to post on our platform. Community notes might be really fascinating there. The interaction with Community notes is awesome. Let's hope it holds up.

在这样一个分裂且充斥着大量病毒式传播帖子的政治气候中，社区笔记似乎是一股必要的新鲜空气。它很棒。没有系统是完美的，但社区笔记的成功率非常高。我还没见过一个错误的笔记能存活超过几个小时。你如何解释它为什么有效？

In a political climate so divisive with intensely viral posts, Community notes seems like an essential breath of fresh air. It's great. No system is perfect, but the batting average of Community notes is incredibly good. I've yet to see an incorrect note that survived more than a few hours. How do you explain why it works?

社区笔记的魔力在于，它要求历史上在笔记评分上存在分歧的人达成一致。要撰写或评分一条笔记，你必须先评分很多笔记。我们使用 AI 来构建一个向量空间，围绕一个人过去如何评分笔记。这并不像左派或右派那么简单；生活要复杂得多。你如何评分社区笔记帖子有很多相关性。要让一条社区笔记显示出来，历史上在该主题上存在分歧的人必须达成一致。这就是它的核心魔力。

The magic of Community notes is that it requires people who have historically disagreed in how they've rated notes to agree. To write or rate a note, you have to rate many notes. We use AI to populate a vector space around how someone has rated notes in the past. It's not as simple as left or right; life is much more complex. There are many correlations in how you rate Community notes posts. For a Community note to be shown, people who historically disagreed on a subject must agree. That's the essential magic of it.

这很迷人，因为有一群存在分歧的人，他们通过这种分歧过程合作，提供背景信息。这说得通，如果过去有分歧的人对某事达成一致，那很可能是真的。我想知道是否会有某种涌现的东西挑战维基百科，维基百科更像是永久性的文章。

It's fascinating because there's a pool of people that have disagreements, and somehow they collaborate through that process of disagreement to come up with context. It makes sense that if people who in the past have disagreed agree about something, it's probably true. I wonder if there's some emergent thing that could challenge Wikipedia, which is more like permanent articles.

维基百科是非常等级制的，而社区笔记本质上不是。没有等级制度。终极考验是，如果有人拿枪指着我的头，我也无法更改一条社区笔记。社区笔记的所有代码都是 100% 开源的，所有数据也是 100% 开源的。你可以独立地完全重建系统中的任何笔记。如果有任何干扰，你会立即注意到。

Wikipedia is very hierarchical, whereas Community notes is inherently not. There is no hierarchy. The acid test is that I can't change a Community note if somebody put a gun to my head. Community notes has all code open source, 100%, and all data open source, 100%. You can completely recreate any note in the system independently. If there was any interference, you'd notice immediately.

顺便说一句，关于 Grok 的话题，它实际上很擅长批评你，这很棒。我问了 Elon Musk 在哪些事情上错了。它基本上严格列出了所有你预测或承诺某件事会发布但实际没有发布的情况。开头语很妙：‘Elon Musk，一个在错误工厂里比错误美国人还错得更多的人。’然后它继续列出。这很有趣。

By the way, on the topic of Grok, it's actually pretty good at criticizing you, which is great. I asked what has Elon Musk been wrong about. It says mostly lists rigorously all the times you predicted or promised when something will be released and it wasn't. The opener is really nice: 'Elon Musk, the man who has been wrong about more things than a wrong American in a wrong factory.' Then it goes on to list. That's pretty funny.

我能看看吗？我以前没见过这个。这很酷。

Can I see it? I haven't seen this before. This is cool.

当然。它说：‘Elon Musk 在哪些事情上错了？嗯，从何说起呢？首先，让我们谈谈臭名昭著的“资金已到位”推文。这和我之前得到的不一样，之前是一个要点列表。这里更有叙事结构。Cybertruck 发布会，他声称牢不可破的窗户……’

Sure. It says: 'What has Elon Musk been wrong about? Well, where do I begin? First off, let's talk about the infamous funding secured tweet. This is different than what it gave me before, which was a bullet point list. Here's more narrative structure. Cyber truck unveiling where he supposedly unbreakable windows...'

哪部分不正确？

Which part is not correct?

第一部分。我确实有资金到位。在旧金山有一场大型民事审判，陪审团一致裁定我无罪。这里暗示实际上资金并未到位。我认为这是从媒体那里来的。那是不正确的。我同意向 SEC 支付罚款的原因，并不是因为 SEC 是正确的。那是 SEC 极其恶劣的行为，坦率地说是腐败。如果我不同意支付罚款，特斯拉会立即破产。我们的 CFO 告诉我，银行会立即暂停我们的信贷额度，如果那样，我们会立刻破产。所以永远不会有审判的机会。会直接死掉。这就像有人拿枪指着你孩子的头说，支付 2000 万美元并承认这一点。就像人质谈判。这个故事被完整讲述了吗？SEC 最好的形式可以成为一股向善的力量，它应该如此，但 SEC 从未追查过任何持续做空并扭曲特斯拉的对冲基金。一次也没有。他们为了自己的利益在电视上公然撒谎，牺牲散户投资者。一次也没有，实际上上千次，SEC 从未追查过他们。你如何解释这种失败？激励机制被破坏了。SEC 的律师薪水不高；这是一份相当低薪的工作。但他们想要的是从 SEC 获得一个奖杯，放在他们的 LinkedIn 上，然后他们可以凭此在高薪律师事务所找到工作。这正是这里的律师所做的。他们不攻击对冲基金的原因，是因为那些对冲基金雇佣了那些律师事务所，他们知道如果攻击对冲基金，会影响他们未来的职业前景。所以他们为了自己的职业生涯牺牲了小投资者。这就是实际发生的事情。监管俘获。不好。所以我接受这件事的唯一原因，技术上是一个‘既不承认也不否认有罪’的和解。但我同意它的唯一原因，是因为我被告知否则特斯拉会破产。如果 SEC 进行这样的调查，银行会暂停资金，我们会立即破产。现在我们处于一个更强大的位置。

The first part. I did actually have the funding secured. There was a big civil trial in San Francisco, and the jury found me not guilty, unanimous finding. Here it's kind of implying that it was not in fact secured. I think this is taking things from the press. That is not correct. The reason I agreed to the fine for the SEC is not because the SEC was correct. That was extremely bad behavior by the SEC, corruption frankly. If I did not agree to pay the fine, Tesla would have gone bankrupt immediately. I was told by our CFO that the banks would immediately suspend our lines of credit, and if they did, we would have gone bankrupt instantly. So there would never have been an opportunity for a trial. It would be dead. It was like someone holding a gun to your kid's head and saying pay $20 million and admit this. Like a hostage negotiation. Was that story fully told? The SEC in its best form could be a force for good, it should be, but not once did the SEC go after any of the hedge funds who were non-stop shorting and distorting Tesla. Not once. They would lie flat out on TV for their own gain at the expense of retail investors. Not once, literally a thousand times, not once did the SEC pursue them. How do you explain this failure? The incentive structure is messed up. The lawyers of the SEC are not paid well; it's a fairly low paying job. But what they're looking for is a trophy from the SEC, something they put on their LinkedIn, from which they can get a job at a high-paying law firm. That's exactly what the lawyer here did. The reason they don't attack the hedge funds is because those hedge funds employ those law firms, and they know if they attack the hedge funds, they're affecting their future career prospects. So they sell small investors down the river for their own career. That's what actually happens. Regulatory capture. Not good. So the only reason I accepted the thing technically was a 'neither admit nor deny guilt' settlement. But the only reason I agreed to that at all was because I was told Tesla would be bankrupt otherwise. If there was an SEC investigation like this, banks would suspend funding, we'd be bankrupt immediately. Now we're in a much stronger position.

听到了吧，Grok。

Take that, Grok.

是的，不幸的是，Grok 从传统媒体那里获取了太多信息。另外，那个人不是洞穴潜水员。有一次，Elon 在一位洞穴潜水员批评他救援被困泰国洞穴男孩的计划后，称他为‘恋童癖’。那个小小的侮辱让他又吃了一场官司，他不得不道歉并支付和解金。那是假的。没有和解。

Yes, unfortunately Grok is taking too much from the conventional media. Also, that guy was not a cave diver. There's a time where Elon called the British cave diver a 'pedo guy' after the diver criticized Musk's plan to rescue a group of boys trapped in a Thai cave. That little slur earned him another lawsuit and he had to apologize and pay a settlement. That is false. There was no settlement.

有个案子，一个不是洞穴潜水员、也不属于救援队的人起诉我，结果输了，一分钱都没拿到。所以这个案子是错的。而且，我想这也是从传统媒体那里来的。实际上，这里有个有趣的问题。这些都是公开的法庭案件。美国证券交易委员会的民事案，民事起诉方输了，旧金山陪审团一致裁决。他们选旧金山是因为觉得那里我最可能输，结果一致裁决对我有利。洛杉矶的审判也是，他们选那个地点是因为觉得我最可能输，又是一致裁决对我有利。两个案子我都赢了。

There was a court case where a guy who is not a cave diver and was not part of the rescue team filed a lawsuit against me and lost. He received nothing. So in this case, it is wrong. It is also, I guess, taking this from the conventional media. Actually, there's an interesting question here. These are public court cases. Both the SEC civil case, where the civil complaints on the SEC guys lost, unanimous jury verdict in San Francisco. They picked San Francisco because they thought it was the place I was most likely to lose, and a unanimous verdict in my favor. The LA trial was also, they picked that venue because they thought I was most likely to lose, unanimous verdict in my favor. Both cases, I won.

对，这里有个有趣的问题。如果新闻机构写一篇关于你埃隆·马斯克的负面文章，似乎能获得更多点击量。这简直是获取点击的最佳方式之一。那么，如果你在训练 Grok，如何避免训练那些激励错位的文章？我们需要把实际法律判决加入训练集。

Yeah, there's an interesting question here. There seems to be a lot more clicks if a journalistic organization writes a negative article about you, Elon Musk. That's like one of the best ways to get clicks. So how do you, if you're training Grok, not train on articles that have misaligned incentives? We need to add the training set of the actual legal decisions.

如果是这样，那是个提示。这实际上很有帮助，因为如果你真的去读法庭结论——这些是公开的——它们和媒体写的完全相反。所以总是追求事实真相。对，除了法官实际写的，陪审团和法官到底得出了什么结论？两个案子都判我无罪。而且那是在陪审团试图找我最可能输的地点之后。

If so, that is a note. This is actually helpful, because if you actually read the court conclusions, which are public, they're completely the opposite of what the media wrote. So always striving for the ground truth. Yeah, beyond the judge actually write, what do the jury and the judge actually conclude? And in both cases, they found me innocent. And that's after the jury show for trying to find the venue where I'm most likely to lose.

不，我的意思是，这显然可以是一个更严厉的批评。我的意思是，我对 Autopilot 过于乐观了。顺便说一句，我得到的批评更多是关于那个。那就是，它为你每个公司列出了漂亮的要点清单，你做出的预测，什么时候能交付，什么时候能解决比如自动驾驶，然后给你一个清单，还挺有说服力的。基本结论就是你经常对完成某事所需的时间过于乐观。

No, I mean, this is obviously can be a much greater critique than this. I mean, I've been far too optimistic about Autopilot. That was the critique I got, by the way, was more about that. Which is, it for each broke down a nice bullet point list for each of your companies, the set of predictions that you made when you will deliver, when you'll be able to solve for example self-driving, and gives you like a list, and it kind of compelling. And the basic takeaway is like you're often too optimistic about how long it takes to get something done.

是的，我是说，我会说我在时间表上病态地乐观。这是真的。但虽然我有时会迟到，我最终总能交付。除了 Uber Lilith？不，我们等着瞧。

Yeah, I mean, I would say that I'm pathologically optimistic on schedule. This is true. But while I am sometimes late, I always deliver in the end. Except with Uber Lilith? No, we'll see.

好的。在过去一年左右，自从收购 X 以来，你变得更加政治化。你有没有一部分后悔那样做？

Okay. Is there, over the past year or so since purchasing X, you've become more political. Is there a part of you that regrets that?

我在这场战斗中是为了对抗觉醒病毒？是的，我想如果你认为对抗觉醒病毒——我认为这是文明威胁——是政治性的，那么是的。

Have I in this battle to sort of counter the woke mind virus? Yeah, I guess if you consider fighting the woke mind virus, which I consider to be a civilizational threat, to be political, then yes.

所以基本上进入政治战场，你有没有一部分后悔？

So basically going into the battleground of politics, is there a part of you that regrets that?

是的，我不知道这一定是支持某个候选人还是另一个，但它通常反对反精英主义的东西，或者试图压制讨论的地方，甚至不允许讨论某个话题。觉醒病毒是共产主义改头换面。

Yes, I don't know if this is necessarily sort of one candidate or another candidate, but it's generally against things that are anti-meritocratic, or where there's an attempt to suppress discussion, where even discussing a topic is not allowed. The woke mind virus is communism rebranded.

嗯，我是说，正因为对抗觉醒病毒的战斗，你被视为右翼。如果觉醒是左翼，那么我想那是对的。但我不确定。我认为左翼有些方面是好的。我是说，如果你支持环境，或者你想为人类创造积极的未来，如果你相信对同胞的同情，善良或不残忍，无论那些价值观是什么。你之前说过你曾是左派或中间偏左。你希望看到什么才能再次投票给民主党？

Well, I mean, that's said because of that battle against the woke mind virus, you're perceived as being right-wing. If the woke is left, then I suppose that would be true. But I'm not sure. I think there are aspects of the left that are good. I mean, if you're in favor of the environment, or if you want to have a positive future for humanity, if you believe in empathy for your fellow human beings, being kind or not cruel, whatever those values are. You said that you were previously left or center-left. What would you like to see in order for you to vote for Democrats again?

不，我会说我在社会问题上可能偏左，在经济问题上可能稍微偏右，这仍然成立。是的。但我想那大概占全国一半，不是吗？也许更多。

No, I would say that I would be probably left of center on social issues, probably a little bit right of center on economic issues, and that still holds true. Yes. But I think that's probably half the country, isn't it? Maybe more.

你和 AOC 是秘密朋友吗？或者更大的问题，你希望你和她的关系，以及所有政治派别的人，能更有同理心地交谈，也许在网上多一点乐趣、好氛围和幽默？

Are you and AOC secretly friends? Or bigger question, do you wish you and her, and just people in general of all political persuasions, would talk more with empathy and maybe have a little bit more fun and good vibes and humor online?

我总是支持幽默。所以我们才有搞笑模式。但好氛围、友情、幽默、友谊。是的。我不认识 AOC。我在 Met Gala 上见过她，她穿着一条裙子。但我只能看到一面，所以看起来像是“吃富人”。我不知道其余部分写了什么。

I'm always in favor of humor. That's why we have funny mode. But good vibes, camaraderie, humor, friendship. Yeah. I don't know AOC. I was at the Met Gala when she attended, and she was wearing this dress. But I could only see one side of it, so it looked like "Eat the Rich." I don't know what the rest of it said.

对，不太确定。跟我说说“吃富人”。我觉得我们应该让语言模型补全那个句子的可能方式。所以我想那不太顺利。好吧，还有希望。我支持友谊。当然，听起来不错。更卡罗尔式。

Yeah, something not sure. Tell me about the "Eat the Rich." I think we should have a language model complete what are the possible ways to complete that sentence. And so I guess that didn't work out well. Well, there's still hope. I root for friendship. Sure, sounds good. More Carolistic.

你是世界上最有名、最富有、最有权势的人之一，如果不是最的话。在你的位置上，很难找到可以信任的人。

You're one of, if not the most, famous, wealthy, and powerful people in the world. In your position, it's difficult to find people you can trust.

不要相信任何人。连自己都不要信。

Trust no one. Not even yourself.

不信任自己？好吧，你是在开玩笑，但有没有“不要相信任何人，连自己都不要信”的成分？没人会花一个小时去想那个，也许还要点毒品，也许还有 Grok。我是说，当生活在一个每个人都想从你这里得到什么的世界里，有没有一些方面？在那个世界里生存有多难？

Not trusting yourself? Okay, well, you're saying that jokingly, but is there some "trust no one, not even yourself"? No one going, need an hour just to think about that, and maybe some drugs, and maybe Grok. I mean, is there some aspect to that when just existing in a world where everybody wants something from you? How hard is it to exist in that world?

我会活下去。有首歌叫《我会活下去》。你一开始会石化吗？好吧，我忘了剩下的歌词。但有没有，你不为此挣扎？我是说，我知道你活下来了，但就像，有办法。石化作为法术，德鲁伊树。它做什么？石化，把怪物变成石头。哦，像字面意思？对，大概六秒。好吧，第二点，暗黑破坏神里有太多数学，让我脑子炸了。就像不停的数学。我是说，真的，你在嘲笑它，但你没有。它会给大脑带来巨大压力。是的，有时肯定很有压力。

I'll survive. There's a song like that too, "I Will Survive." Will you petrified at first? Okay, I forget the rest of the lyrics. But is there, you don't struggle with this? I mean, I know you survive, but like, there's ways. Petrify as a spell, the Druid tree. What does it do? Petrify, it turns the monsters into stone. Oh, like literally? Yeah, for like six seconds. Well, the second there's so much math in Diablo that breaks my brain. It's like math nonstop. I mean, really, you're like laughing at it, but you don't. It can put a huge amount of tension on a mind. Yes, it can be definitely stressful at times.

那么，你怎么知道在工作和个人生活中可以信任谁？

Well, how do you know who you can trust in work and personal life?

我是说，我想你会看一个人长期的表现。如果他们有一个，你知道，我想你有点像用你的神经网络来评估一个人。神经网络不会感到痛苦。你的神经网络有意识，当人们背叛你时，它可能会感到痛苦。可以让人，我是说，坦白说，我几乎从未被背叛过。非常非常罕见。所以你知道，不管怎样，我想因果报应，对人好，他们也会对你好。是的，因果报应是真实的。

I mean, I guess you look at somebody's track record over time. And if they've got a, you know, I guess you kind of use your neural net to assess someone. Neural nets don't feel pain. Your neural net has consciousness, it might feel pain when people betray you. Can make, I mean, to be frank, I've almost never been betrayed. It's very, very rare. So you know, for what it's worth, I guess karma, be good to people and they'll be good to you. Yeah, karma is real.

有没有你信任的人？让我修改一下这个问题。有没有亲近的人说你古怪？还有你的，嗯，专家形式对此很有帮助，如果你在寻找批评性反馈。它会不会把你推向更极端，思想的极端，让你对人性普遍变得愤世嫉俗？

Are there people you trust? Let me edit that question. Are there people close to you that call you odd? And your, well, the expert form is very helpful for that if you're looking for critical feedback. Can it push you into the extremes more, the extremes of thought, make you cynical about human nature in general?

我不认为我会。

I don't think I will be.

事实上，我认为永远不要相信一个愤世嫉俗的人。原因是愤世嫉俗者会以“大家都这么做”来为自己的不良行为开脱。所以如果有人是真正的愤世嫉俗者，我会认为这是一个危险信号。这其中有一定程度的投射，从外部看这种虚伪很有趣。但有一点很重要：如果一个人愤世嫉俗，认为所有人都有不良行为，他们就很容易用“大家都这么做”来为自己的行为开脱。这不是真的。大多数人都是中等偏好的。我希望 X 平台上的人能更善于看到他人行为中的优点。似乎存在一种偏向负面的倾向；负面信息更性感、更容易传播。我不确定这是人性的哪一部分。

In fact, I think one should never trust a cynic. The reason is that cynics excuse their own bad behavior by saying everyone does it. So I always see it as a red flag if someone is a true cynic. There's a degree of projection there, which is fun to watch from the outside and enjoy the hypocrisy. But this is an important point: if somebody is cynical, meaning they see bad behavior in everyone, it's easy for them to excuse their own bad behavior by saying everyone does it. That's not true. Most people are kind of medium good. I do wish people on X would be better at seeing the good in others' behavior. There seems to be a bias towards seeing the negative; the negative is sexier, more viral. I don't know what that is about human nature.

我发现 X 平台比传统媒体负面情绪更少。如果你读传统报纸，只会让你难过。而在 X 上，我每天得到的笑声比所有其他东西加起来还多。

I find the X platform to be less negative than legacy media. If you read a conventional newspaper, it just makes you sad. Whereas on X, I really get more laughs per day than everything else combined.

笑是一回事，但它与积极氛围和支持并不完全重叠，比如以很棒的方式庆祝他人。感觉负面情绪占了上风。但这比主流媒体好，主流媒体几乎对一切都无情地负面。传统新闻试图回答：今天地球上发生了什么最糟糕的事？世界很大，所以每天都有坏事发生。新闻中就是有强烈的负面偏见。一个可能的解释是进化：历史上坏消息可能是致命的，比如狮子或敌对部落。好消息，比如找到浆果，虽然好但不是必需的。

Laughs are one thing, but they don't perfectly overlap with good vibes and support, like celebrating others in an awesome way. It feels that is outweighed by negativity. But it's better than mainstream media, which is almost relentlessly negative about everything. Conventional news tries to answer: what is the worst thing that happened on Earth today? And it's a big world, so on any given day something bad has happened. There's just a strong negative bias in the news. A possible explanation is evolutionary: bad news historically could be fatal, like a lion or an enemy tribe. Good news, like finding berries, is nice but not essential.

我们的老朋友特斯拉自动驾驶可能是世界上最智能的现实世界 AI 系统之一。你从一开始就关注它。它一直是最不可思议的机器人之一。当它从轮式机器人泛化为一个感知世界并可能拥有不同形态的现实世界 AI 系统时，非常令人兴奋。

Our old friend Tesla Autopilot is probably one of the most intelligent real-world AI systems in the world. You followed it from the beginning. It was one of the most incredible robots and continues to be. It was exciting when it generalized from a robot on wheels to a real-world AI system that perceives the world and can have different embodiments.

端到端训练真正神奇的地方在于它学会了阅读，比如你可以阅读科学文章，但我们从未教过它阅读。我们从未教过它什么是汽车、人或骑自行车的人。它从视频中学会了路上所有物体是什么，仅仅通过观看视频。就像人类一样：光子输入，控制输出。到达大脑的绝大多数信息来自眼睛。输出是我们手指和嘴巴的运动信号。光子输入，控制输出。汽车也是如此。通过观察图像序列，它找到相关聚类并理解世界的深层含义。这很美。我们的大脑就是这样工作的。问题是它能走多远？LLM 在文本上进行自监督学习与特斯拉自动驾驶似乎有深刻的相似性。它们都朝着 AGI 发展，但特斯拉的方法在计算效率上高得多，因为它不得不如此。我们只有 100 瓦特和每秒 144 万亿次运算的大脑计算机。它只用 256 个值理解世界。通过现实世界理解走向 AGI 的路径可能影响更大，因为它会比 LLM 更快地理解现实世界并更快地与人类融合。特斯拉的方法从根本上更高效。我们的大脑非常高效节能。大脑的思考部分不到 10 瓦特，却能写出比 10 兆瓦 GPU 集群更好的小说。这是六个数量级的差异。AI 通过蛮力达到现在的水平，但这不会是终点。随着时间的推移，模型会变小，用更少的算力和功耗产生合理的输出。特斯拉在这方面可以说领先，因为我们被迫用 100 瓦特的算力理解世界。我们用硬件 3 和硬件 4 修复了许多基本功能，硬件 5 会更好。看起来即使使用硬件 3 和 100 瓦特，汽车也能比人类驾驶得更好，经过优化可能不到 50 瓦特。

The really wild thing about the end-to-end training is that it learns to read, like you can read science, but we never taught it to read. We never taught it what a car was, or a person, or a bicyclist. It learned what all those objects on the road are from video, just from watching videos. Just like humans: photons in, controls out. The vast majority of information reaching our brain is from our eyes. The output is our motor signals to our fingers and mouth. Photons in, controls out. The same is true of the car. By looking at the sequence of images, it finds correlative clusters and understands something deep about the world. It's beautiful. That's how our brain works. The question is how far can it go? There seems to be a deep similarity between what LLMs do with self-supervised learning on text and what Tesla Autopilot does. They are both headed towards AGI, but Tesla's approach is much more compute-efficient because it had to be. We only have 100 watts and a brain computer of 144 trillion operations per second. It's understanding the world with only 256 values. The path to AGI through real-world understanding might have more significant impact because it will understand the real world faster than LLMs and integrate with humans faster. Tesla's approach is fundamentally more compute-efficient. Our brain is very compute-efficient and energy-efficient. The thinking part of our brain is less than 10 watts, yet it can produce a much better novel than a 10-megawatt GPU cluster. That's a six orders of magnitude difference. AI has gotten to where it is via brute force, but that's not where it will end up. Over time, models will get smaller and produce sensible output with far less compute and power. Tesla is arguably ahead on that front because we've been forced to understand the world with 100 watts of compute. We fixed a bunch of fundamental functions with Hardware 3 and Hardware 4, and Hardware 5 will be even better. It appears that the car will be able to drive better than a human even with Hardware 3 and 100 watts, and with optimization, probably less than 50 watts.

你在开发 Optimus 以及将这种现实世界 AI 应用于机器人操作和人形机器人方面学到了什么？

What have you learned about developing Optimus and applying this real-world AI to robotic manipulation and humanoid robotics?

你理解到的一些有趣的小事或大事。我惊讶于我们必须自己开发机器人的每一个部件，没有现成的电机、电子设备、传感器。我们不得不开发一切。

Some interesting tiny or big things you've understood. I was surprised at the fact that we had to develop every part of the robot ourselves, that there were no off-the-shelf motors, electronics, sensors. Like we had to develop everything.

我们实际上根本找不到任何价位的电机来源。所以这不仅仅是效率或成本的问题，而是根本没有。实际上，一切都必须从头设计。我们努力寻找任何现成的东西，因为你想世界上制造了多少电机，有成千上万种电机设计。但没有一种适合人形机器人，一个都没有。所以我们不得不开发自己的设计，专门针对人形机器人的需求来设计。

We couldn't actually find a source of electric motors for any amount of money. So it's not even just efficient or inexpensive, it's like anything. There's nothing. Actually, everything has to be designed from scratch. We tried hard to find anything that was... because you think of how many electric motors are made in the world, there are tens of thousands, hundreds of thousands of electric motor designs. None of them were suitable for a humanoid robot, literally none. So we had to develop our own design, design it specifically for what a humanoid robot needs.

设计一个可以大规模生产、相对便宜的机器人有多难？我的意思是，如果和波士顿动力的 Atlas 相比，它非常昂贵。

How hard was it to design something that can be mass manufactured, could be relatively inexpensive? I mean, if you compare to Boston Dynamics Atlas, it's a very expensive robot.

它的设计是为了像制造汽车一样进行生产。我认为最终我们可以让 Optimus 的成本低于一辆汽车。应该如此，因为你看机器人的质量要小得多，而汽车有很多执行器。汽车的执行器比机器人多，但人形机器人手指的执行器很有意思。所以 Optimus 有非常灵巧的手和手指，它们可以进行一些有趣的操作，比如软触机器人技术。

It is designed to be manufactured in the same way that we would make a car. And I think ultimately we can make Optimus for less than the cost of a car. It should be, because if you look at the mass of the robot, it's much smaller, and the car has many actuators in it. The car has more actuators than the robot, but the actuators are kind of interesting in a humanoid robot with fingers. So Optimus has really nice hands and fingers, and they can do some interesting manipulation, soft touch robotics.

我的一个测试目标，是它能不能拿起针和线，仅通过视觉穿针引线？我们离那还有多远？仅通过视觉。

One of the tests, goals I have, is can it pick up a needle and a thread and thread the needle just by looking? How far away are we from that? Just by looking.

也许一年。尽管我回过头来，我对时间持乐观态度。我们在汽车上做的工作会迁移到机器人上。感知，或者控制。不，控制不同，但视频输入、控制输出。汽车是四个轮子的机器人。Optimus 是有手有脚的机器人。所以它们非常相似。所以整个学习过程的机制，你只是有一套不同的控制。Optimus 会通过观看视频来学会做事。

Maybe a year. Although I go back to, I'm optimistic on time. The work that we're doing in the car will translate to the robot. The perception, or also the control. No, the controls are different, but the video in, controls out. The car is a robot on four wheels. The Optimus is a robot with hands and legs. So they're very similar. So the entire machinery of the learning process, you just have a different set of controls. Optimus will figure out how to do things by watching videos.

俗话说，‘善待他人，因为你遇到的每个人都在打一场你一无所知的仗。’这是真的。你正在经历什么人们通常看不到的困难？

As the saying goes, 'Be kind, for everyone you meet is fighting a battle you know nothing about.' It's true. What's something difficult you're going through that people don't often see?

试图喂饱 Uber？不，我是说，我的思绪像风暴一样，我不认为大多数人想成为我。他们可能以为他们想成为我，但他们不知道，他们不理解。

Trying to feed Uber? No, I mean, you know, my mind is a storm, and I don't think most people would want to be me. They may think they would want to be me, but they don't know, they don't understand.

你怎么样？我是说，总体上还好吗？

How are you doing? I mean, overall okay?

从大局来看，我没什么可抱怨的。

In the grand scheme of things, I can't complain.

你有时会感到孤独吗？

Do you get lonely sometimes?

但我的孩子和朋友陪伴着我，所以不是存在性的孤独。但有很多夜晚我独自入睡。我不必这样，但我确实如此。

But my kids and friends keep me company, so not existential. But there are many nights I sleep alone. I don't have to, but I do.

沃尔特·艾萨克森在你新的传记中写到了你艰难的童年。你会在心中原谅你生命中那段时期发生的一切吗？

Walter Isaacson in his new biography of you wrote about your difficult childhood. Will you ever find forgiveness in your heart for everything that has happened to you in that period of your life?

什么是原谅？我没有，至少我不认为我有怨恨，所以没什么可原谅的。你知道，原谅对人们来说很难。看起来你没有心怀怨恨。

What is forgiveness? I do not, at least I don't think I have a resentment, so nothing to forgive. You know, forgiveness is difficult for people. It seems like you don't harbor the resentment.

我的意思是，我试着思考什么会以好的方式影响未来，而心怀怨恨不会以好的方式影响未来。

I mean, I try to think about what is going to affect the future in a good way, and holding on to grudges does not affect the future in a good way.

你是一位父亲，一位自豪的父亲。你从你的孩子身上学到了什么关于生活的东西？

You're a father, a proud father. What have you learned about life from your kids?

那些小的生物有机体。我的意思是，开发 AI 和看着，比如说，小 X 成长是迷人的，因为相似之处比我预期的要多得多。我是说，我可以看到他的生物神经网络越来越理解世界，同时我也看到数字神经网络越来越理解世界。

Those little biological organisms. I mean, developing AI and watching, say, a little X grow is fascinating, because there are far more parallels than I would have expected. I mean, I can see his biological neural net making more and more sense of the world, and I can see the digital neural net making more and more sense of the world at the same time.

你在两者中都看到了美和魔力吗？

Do you see the beauty and magic in both?

是的。我的意思是，和孩子在一起的一件事是，你通过他们的眼睛重新看待世界。对他们来说，一切都是新鲜和崭新的，当你看到他们体验世界时，世界对你来说也是新鲜和崭新的。

Yes. I mean, one of the things with kids is that you kind of see the world anew in their eyes. To them, everything is new and fresh, and when you see them experience the world, the world is new and fresh to you too.

好了，埃隆，我只想说感谢你多年来对我的善意和友谊，感谢你在我这样一个傻孩子身上看到了什么，就像你为许多其他人所做的那样，感谢你对人类积极未来抱有希望，并为此拼命工作。谢谢。

Well, Elon, I just want to say thank you for your kindness to me and friendship over the years, for seeing something in a silly kid like me, as you've done for many others, and thank you for having hope for a positive future for humanity and for working your ass off to make it happen. Thank you.

谢谢，莱克斯。

Thanks, Lex.

感谢收听与埃隆·马斯克的这次对话。要支持本播客，请查看描述中的赞助商。现在让我用沃尔特·艾萨克森写的关于埃隆处理困难问题的核心哲学的话来结束：‘唯一的规则是由物理定律决定的规则。’感谢收听，希望下次再见。

Thank you for listening to this conversation with Elon Musk. To support this podcast, please check out our sponsors in the description. And now let me leave you with some words that Walter Isaacson wrote about the central philosophy of how Elon approaches difficult problems: 'The only rules are the ones dictated by the laws of physics.' Thank you for listening, and hope to see you next time.