【双语音+文稿】（完整版）英伟达创始人兼CEO黄仁勋10月最
共 1 集
剧集目录
| 集 | 标题 | 时长 |
|---|---|---|
| P1 | 【双语音+文稿】（完整版）英伟达创始人兼CEO黄仁勋10月最 | 1:39:56 |
P1 · 【双语音+文稿】（完整版）英伟达创始人兼CEO黄仁勋10月最新访谈：内容涵盖AI计算需求激增、技术突破、安全规范、产业生态构建及公司战略等。 p01 英文版 (P1)
【双语音+文稿】（完整版）英伟达创始人兼CEO黄仁勋10月最新访谈：内容涵盖AI计算需求激增、技术突破、安全规范、产业生态构建及公司战略等。 p01 英文版 (P1)
发布于 2026-10-17 · 时长 1:39:56字幕摘录
| 时间 | 英文 | 中文 |
|---|---|---|
| 0:00 | This is an existential issue for us. | 这对我们来说是一个存在的问题。 |
| 0:02 | We could easily stop, but if we stop today, we lose. | 我们可以轻易停下来,但如果我们今天停下来,我们就输了。 |
| 0:05 | Functionality, capability, safety, security, they're all at exactly the same place. | 功能性,能力,安全性,安全性,都完全相同. |
| 0:10 | Priority number one. | 首要任务 |
| 0:21 | It's our third pod back together. | 这是我们的第三舱一起回来。 |
| 0:23 | In the last 10 days, you've been at the Shee State Dinner at the White House. | 过去10天里,你一直在白宫的谢州晚餐会上. |
| 0:28 | You were at the White House Summit where we did the accord on super intelligence. | 你当时在白宫峰会上 我们达成了超级情报协议 |
| 0:33 | Just yesterday, you were there with your parents | 就在昨天,你和父母一起在那边 |
| 0:36 | when the president presented you with the National Medal of Science. | 当总统给你颁发国家科学奖章时 |
| 0:40 | You were the only investor at the table. | 你是唯一的投资人 在桌子上。 |
展开字幕全文（1945 条）
| 序号 | 英文 | 中文 |
|---|---|---|
| 1 | This is an existential issue for us. | 这对我们来说是一个存在的问题。 |
| 2 | We could easily stop, but if we stop today, we lose. | 我们可以轻易停下来,但如果我们今天停下来,我们就输了。 |
| 3 | Functionality, capability, safety, security, they're all at exactly the same place. | 功能性,能力,安全性,安全性,都完全相同. |
| 4 | Priority number one. | 首要任务 |
| 5 | It's our third pod back together. | 这是我们的第三舱一起回来。 |
| 6 | In the last 10 days, you've been at the Shee State Dinner at the White House. | 过去10天里,你一直在白宫的谢州晚餐会上. |
| 7 | You were at the White House Summit where we did the accord on super intelligence. | 你当时在白宫峰会上 我们达成了超级情报协议 |
| 8 | Just yesterday, you were there with your parents | 就在昨天,你和父母一起在那边 |
| 9 | when the president presented you with the National Medal of Science. | 当总统给你颁发国家科学奖章时 |
| 10 | You were the only investor at the table. | 你是唯一的投资人 在桌子上。 |
| 11 | Well, now that we reflected on it. | 瞷и稱筁 |
| 12 | Yeah. | 对 |
| 13 | Well, tell me, being there with your parents yesterday, | 告诉我,昨天跟你父母在一起 |
| 14 | it struck me five out of the six people on stage were all born outside of the United States. | 舞台上的六个人中有五个出生在美国以外 |
| 15 | How did that feel? | 感觉怎么样? |
| 16 | Well, first of all, it's a great honor all by itself. | 首先,这本身就是一个伟大的荣誉。 |
| 17 | It's a reflection of the contribution that the company has made. | 反映公司的贡献. |
| 18 | Obviously, we created, we invented a new computing model. | 显然,我们创造了一个新的计算模型。 |
| 19 | This is probably the most pervasive computing model in the world today. | 这可能是当今世界最普遍的计算模型。 |
| 20 | It's the only computing architecture that's in every cloud. | 这是每个云层中唯一的计算架构. |
| 21 | It's in every computer company. | 在每个电脑公司都有 |
| 22 | It's used in every field of science. | 用于科学的每一个领域. |
| 23 | It's used by gamers. | 由游戏者使用. |
| 24 | It's used by creators and builders and it's used by scientists. | 它被创作者和建筑者使用,被科学家使用. |
| 25 | It powers intelligence. | 它能赋予智慧 |
| 26 | If you take a step back and you look at what the company has accomplished, | 如果你退一步 看看公司的成就 |
| 27 | we started out building a graphics chip for gamers. | 我们开始为玩家建造图形芯片 |
| 28 | I'm still proud of that, of course. | 我当然还是为此感到骄傲 |
| 29 | We now shipped one and a half, 1.67 billion units. | 我们现在运送了一个半, 16.7亿单位。 |
| 30 | I raised generations of scientists and kids and designers. | 我培养了几代科学家、孩子和设计师。 |
| 31 | So, I think the impact, and here we are, we're the infrastructure | 因此,我认为影响, 在这里,我们是基础设施 |
| 32 | that powers the intelligence of society. | 社会智慧的力量 |
| 33 | It's an incredibly proud moment for me to receive the award on behalf of the company. | 代表公司获得奖项 对我来说真是令人骄傲的时刻 |
| 34 | Now, for my parents, I wanted them to go and I gave them literally | 现在,为了我的父母, 我希望他们去, 我给他们字面上 |
| 35 | 24 hours of time and they don't travel that much. | 24小时的时间,他们不 旅行那么多。 |
| 36 | They don't travel that much. | 他们旅行不多。 |
| 37 | My dad didn't even have a suit. | 我爸爸甚至没有西装。 |
| 38 | Wow. | 珃 |
| 39 | Yeah. | 对 |
| 40 | So, he ran out and bought a suit. | 所以他跑出去买了一套西装 |
| 41 | He looked great. | 他看起来很棒 |
| 42 | They looked great. | 他们看起来很棒。 |
| 43 | Yeah. | 对 |
| 44 | It was a brand new suit and he showed up. | 这是全新的西装,他出现了。 |
| 45 | I didn't realize he had a suit. | 我不知道他有西装 |
| 46 | I thought he was going to wear some shirt. | 我以为他会穿件衬衫 |
| 47 | He went out and bought a suit. | 他出去买了套西装 |
| 48 | My mom got a new dress and literally in 24 hours. | 我妈妈在24小时内换了件新衣服 |
| 49 | Oh, I know how it works. | 哦,我知道如何运作。 |
| 50 | Oh my gosh. | 天呐 |
| 51 | It's incredible. | 不可思议. |
| 52 | We chartered a plane for them, got them there. | 我们为他们包了一架飞机 把他们送到那里了 |
| 53 | And they were so happy to meet the president and President Trump walked my mom because | 他们很高兴见到总统 特朗普总统陪着我妈妈 |
| 54 | she has a hard time walking and the two of them were walking off stage and he held her | 她走不动,他们两个走下台 他抱着她 |
| 55 | and walked her down the stage and they were just incredibly, incredibly proud to meet her. | 把她带到舞台上 他们非常骄傲地遇见了她 |
| 56 | But for them, it's their American dream. | 但对他们来说,这是他们的美国梦. |
| 57 | My dad, I still remember, I think I was like five years old or something like that and | 我爸爸,我还记得, 我想我像5岁什么的, |
| 58 | he had just come back from the United States on a business trip. | 他刚从美国回来出差 |
| 59 | He was working for an air conditioning company. | 他在一家空调公司工作 |
| 60 | And you were living in Taiwan. | 你当时住在台湾 |
| 61 | We were at the time, we had just moved to Thailand. | 当时我们刚搬到泰国 |
| 62 | Thailand. | 泰国。 |
| 63 | And no, no, we were just about to move to Thailand and we were in Taiwan. | 不,不,我们正准备搬去泰国 我们在台湾。 |
| 64 | He had come back and he was in the United States working for a company called Carrier. | 他回来后,他在美国为一家叫Carrier的公司工作。 |
| 65 | And he was here for training and of all the places in the world, they landed him in New York. | 他在这里训练 和世界上所有的地方, 他们把他在纽约。 |
| 66 | Okay. | 摆 |
| 67 | And so you can just imagine a kid in the late 20s, I guess, late 20s, and never been outside | 所以你可以想象一个20年代末的孩子, 我想,20年代末, 从来没有在外面 |
| 68 | of Taiwan, goes to United States, New York City of all places and he saw the future of the future. | 台湾,去美国,纽约 各地,他看到了未来。 |
| 69 | And he came back, mind blown. | 他回来了,心碎了 |
| 70 | All the things that he bought at the airport and brought it back, he brought chocolate, | 他在机场买回来的东西 他都带了巧克力 |
| 71 | he brought, you know that cheese that you have in a tube back in the old days? | 他带来了,你知道 你有一个奶酪 在管子的回到过去? |
| 72 | Strain cheese. | 草条奶酪。 |
| 73 | Strain cheese that comes in a tube. | 绳子里的奶酪 |
| 74 | That's great. | 赞曰. |
| 75 | And he brought back all kinds of strange things and he says, this is, we should go there someday. | 他带回了所有奇怪的东西 他说,这是,我们应该 有一天去那里。 |
| 76 | And the first opportunity they had, they sent my older brother and I to United States and | 他们把我和哥哥送到美国 |
| 77 | you know, it was his dream. | 这是他的梦想 |
| 78 | I don't think many people realize that you came to the United States ahead of your parents. | 我不认为很多人意识到你比父母早来到美国. |
| 79 | Yeah. | 对 |
| 80 | Right. | 对 |
| 81 | And they joined a few years later. | 几年后他们也加入了 |
| 82 | But it just hit me like a ton of bricks that five out of the six people on that stage yesterday | 但昨天在舞台上六人中有五个人 把我打得一团糟 |
| 83 | were born outside the United States. | 出生在美国境外。 |
| 84 | You and I have talked about the need for, you know, with all the other stuff going on, | 你和我谈过需要,你知道, 与所有其他的事情, |
| 85 | a real strategy and an immigration policy that makes certain that we're still the place that | 真正的策略和移民政策 确保我们仍然是 |
| 86 | all the great builders in the world want to come build. | 世界所有伟大的建筑家都想来建造. |
| 87 | This is America's superpowers. | 这是美国的超能力。 |
| 88 | Right. | 对 |
| 89 | In a lot of ways, the intellectual property creation capability of our nation | 在许多方面,我们国家的知识产权创造能力 |
| 90 | is unlike anything else in the world. | 和世界上其他的不一样 |
| 91 | And the reason for that is we attract the world's best. | 原因就是我们吸引了全世界最好的。 |
| 92 | We are the place where they want to come and build their lives | 我们就是他们想要来的地方 建设他们的生活 |
| 93 | because of the opportunity it provides, the freedom that she offers, the opportunity that | 因为它提供的机会, 她提供的自由, |
| 94 | it provides, the spirit of invention and entrepreneurialism and the can-do attitude. | 它提供了发明精神和创业精神,以及能行的态度。 |
| 95 | You know, in the United States, everything is possible. | 你知道,在美国,一切都是可能的。 |
| 96 | Right. | 对 |
| 97 | We start with, of course, it's possible. | 我们首先当然可以 |
| 98 | Of course, we could do that. | 当然,我们可以做到这一点。 |
| 99 | Instead of, you know, I do a lot of travel. | 而不是,你知道, 我做很多旅行。 |
| 100 | And of course, you can't do that. | 当然,你不能这样做。 |
| 101 | Of course, you know. | 当然,你知道的。 |
| 102 | In other places. | 在别处. |
| 103 | And we take it for granted. | 我们认为这是理所当然的。 |
| 104 | But the spirit, the energy that comes around with the culture that we've created here | 但是,精神,能量 围绕着我们在这里创造的文化 |
| 105 | just can't be understated. | 只是不能被低估。 |
| 106 | And so I think the flow, the inflow of intellectual property into this country | 所以我认为 知识产权流入这个国家 |
| 107 | is unlike anything else the world's ever seen. | 和世界上见过的不一样 |
| 108 | It's true. | 这是真的。 |
| 109 | And we want that. | 我们希望这样。 |
| 110 | We want all of the world's great minds to come here, bring their great ideas, | 我们希望世界上所有伟大的人才来到这里, 带来他们伟大的想法, |
| 111 | make them, realize them here. | 让他们意识到他们在这里。 |
| 112 | Right. | 对 |
| 113 | Realize them here and let us keep that flywheel going. | 认识他们 让我们继续飞翔 |
| 114 | It's a delicate system. | 微妙系统. |
| 115 | It's a delicate balance I get. | 这是一个微妙的平衡 我得到。 |
| 116 | We have to do it lawfully. | 我们必须合法行事。 |
| 117 | We have to do it rightfully. | 我们必须正确地做到这一点。 |
| 118 | But we surely want to do it. | 但我们确实想这样做。 |
| 119 | Yeah, for sure. | 是的,当然。 |
| 120 | So I want to start high level, just kind of where we are. | 所以我想从高层次开始, 只是那种我们的位置。 |
| 121 | Then I want to dig down a little bit into the safety and regulation and getting this right. | 然后我想深入了解一下安全与监管 并找出问题 |
| 122 | Where we are in terms of the compute build out. | 在计算积分方面我们的位置 |
| 123 | And then finally, we're going to finish up drilling in a little bit to NVIDIA's growth | 最后,我们要完成钻探 一点点NVIDIA的成长 |
| 124 | and what's going on there. | 和发生什么事了。 |
| 125 | But you've been really good at calling the future, right? | 但你一直很擅长召唤未来 对吧? |
| 126 | Two years ago, we were sitting here and you said that the market was | 两年前,我们坐在这里 你说市场是 |
| 127 | dramatically underestimating the amount of compute build out that was going to be required | 大大低估了需要计算的建筑数量 |
| 128 | and was going to happen over the course of the next five years. | 接下来的五年里 将会发生 |
| 129 | You said it's not going to be a trillion. | 你说不会是万亿 |
| 130 | It's going to be two trillion at a minimum in terms of cumulative compute build out. | 累计计算积分至少要2万亿. |
| 131 | And of course, now we're over that. | 当然,现在我们结束了。 |
| 132 | You said before everybody was thinking about it that | 在大家考虑之前你说过 |
| 133 | inference, right, compute was going to one billion X because of inference time reasoning | 推论,对,计算到十亿X 因为推论时间推理 |
| 134 | and thinking. | 和思维。 |
| 135 | And then last year, of course, before we saw Opus 4.6, before we saw Claude Code, | 然后去年,当然, 在我们看到Opus 4.6, 在我们看到克劳德代码之前, |
| 136 | you said we're entering the age of agents, right? | 你说我们要进入特工时代,对吧? |
| 137 | Where we're going to unlock a lot of these capabilities. | 在那里我们要去解锁 很多这些能力。 |
| 138 | 当然,我们看到Opus 4.6和Claude Code真的起飞了。 | |
| 139 | So where are we now, right? | 我们现在在哪? |
| 140 | We have Elon and Sam and Dan. | 我们有伊隆和萨姆 还有丹 |
| 141 | There's only three things that I said that came true. | 我说的是三件事才是真的 |
| 142 | There's another 200. | 还有两百个 |
| 143 | Why don't you just keep on going? | 你为什么不继续走? |
| 144 | Let that be the podcast. | 让播客来播吧 |
| 145 | So our investor day in a couple of weeks is on entering the singularity, this acceleration. | 因此,我们的投资者日 在几周内 进入奇点, 这个加速。 |
| 146 | We have Elon and Sam and Demis all say, hey, we're in the foothills or entering the singularity. | 我们让伊隆和萨姆和戴米斯都说, 嘿,我们在山坡上或者进入奇点。 |
| 147 | We have model-solving Navier-Stokes, as you know. | 我们有模型解 纳维尔 -斯托克斯,你知道的。 |
| 148 | So when you look into your crystal ball, where are we? | 所以当你看着你的水晶球, 我们在哪里? |
| 149 | And are we inflecting? | 我们在反射吗? |
| 150 | Are we accelerating into a different phase of super intelligence? | 我们是否正在加速进入另一个超级智能阶段? |
| 151 | And so let's unpack that. | 所以,让我们打开它。 |
| 152 | First of all, it's fairly logical why the amount of compute that we're going to need | 首先,这是相当合理的 为什么计算的数量 我们需要 |
| 153 | is a billion times more than where we were just recently, let's call it, last year. | 比我们最近那段时间多十亿倍, 让我们称之为去年。 |
| 154 | And the reason for that is because thinking and problem solving takes a lot more inferences than | 原因是思考和解决问题 需要更多的推论 |
| 155 | repeating something, regurgitating something that you memorized. | 重复一些东西, 重现你记忆的东西。 |
| 156 | So when you're a chat GBT, you give it a prompt. | 所以当你是一个聊天的GBT, 你给它一个提示。 |
| 157 | Most of that knowledge was encoded into a large language model. | 大部分知识被编码成一个大语言模型。 |
| 158 | Unfortunately, of course, it hallucinated. | 不幸的是,当然,它产生了幻觉。 |
| 159 | And so the next click was it was generative, but it was also grounded or conditioned by | 接下来的点击是 遗传性的,但它也是基于或条件的 |
| 160 | ground truth. | 事实真相 |
| 161 | It would have to go out and do a search, do a quick search, comes back with the ground truth. | 它必须出去搜身 快速搜身 带着真相回来 |
| 162 | And then from that conditioning, augmented with what it already learned pre-training, | 然后从这个条件, 与它已经学到的预训, |
| 163 | and it would generate whatever the answer is. | 它会产生任何答案。 |
| 164 | Well, most of the things that we want to really do that are interesting requires you to reason | 大部分我们真正想要做的事情 很有趣 需要你解释 |
| 165 | step by step, to come up with several options, plan. | 一步一步地, 提出几个选择,计划。 |
| 166 | And you might decide, let's see how these things all roll out. | 你可能会决定, 让我们看看这些事情是如何展开的。 |
| 167 | And then based on the answers of all of those, I reflect on the best answer so far, | 然后根据所有这些答案, 我思考了迄今为止最好的答案, |
| 168 | the best path so far. | 目前为止最好的路径。 |
| 169 | I might choose one or two or three of them, roll them out, prune the rest of it. | 我可能会选一两三个, 把它们推出来, 将剩下的。 |
| 170 | And then just keep on doing this. | 然后继续做下去 |
| 171 | That's the reason why we say great strategists love strategic optionality. | 这就是为什么我们说伟大的战略家 喜欢战略选择。 |
| 172 | You don't want to prune things too fast on the one hand, but you also have to optimize | 你并不想把事情推得太快 但你也要优化 |
| 173 | your resources. | 你的资源。 |
| 174 | Well, AIs are doing that. | 人工智能正在这样做。 |
| 175 | So they're keeping their strategic optionality. | 所以他们保持了战略上的可选性. |
| 176 | They're rolling things out. | 他们正在把事情搞砸 |
| 177 | run | |
| 178 | They're reflecting. | 他们在反射 |
| 179 | They're pruning, so on and so forth. | 二者普润等. |
| 180 | All of that thinking, all of that exploration just takes a lot more compute. | 所有那些想法,所有探索 只是需要更多的计算。 |
| 181 | And not on top of that, because it's able to now do productive things, we're going to | 更别说了,因为它现在能做生产性的事了, 我们要... |
| 182 | use it for more things. | 用它做更多的东西。 |
| 183 | And so the compounding of both, it uses more compute to do its job, but it's doing a good | 所以两者的复合, 它用更多的计算来完成它的工作, 但是它是一个很好的 |
| 184 | job, and therefore, we're going to hire it to do more jobs. | 因此,我们要雇用它做更多的工作。 |
| 185 | And so that all compounds. | 所以所有的化合物。 |
| 186 | At this point, I would say that the big observation from last year is that on the one hand, we | 此时此刻,我要说,去年的重要观察是,一方面,我们 |
| 187 | want to get better and better, and we still want the models to get better. | 想要越来越好, 我们还希望模型变得更好。 |
| 188 | And the scaling laws of models are going to continue. | 而模型的缩放定律将继续下去。 |
| 189 | However, the attachment of a harness, basically the exoskeleton of the large language model, | 然而,一个带子的附着, 基本上是大语言模型的exoskeleton, |
| 190 | the body of the brain, that allows it to manage working memory, manage knowledge. | 大脑的身体, 允许它管理工作记忆, 管理知识。 |
| 191 | How do you go from working memory to short term, to long term, to basically formulating | 如何从工作记忆到短期、长期、 基本形成 |
| 192 | your knowledge graph, which is what we all do. | 你的知识图,我们都是这么做的。 |
| 193 | We think about, hey, what did I learn, and how does it fit into everything that I know | 我们想着,嘿,我学到了什么, 它如何融入我知道的一切 |
| 194 | in the past? | 过去? |
| 195 | And how when we go to bed, we do compaction. | 当我们上床睡觉的时候,我们做紧凑。 |
| 196 | Things that don't make sense or don't matter, we eventually forget. | 说不通或无关紧要的事情 我们最终会忘记的 |
| 197 | But the things that matter, somehow it gets reconnected. | 但重要的是 不知何故它会重新连接起来 |
| 198 | And so we're doing that offline. | 所以我们要下线了 |
| 199 | And so we've got working memory. | 所以我们有工作记忆。 |
| 200 | We have offline knowledge extraction and creation, knowledge graph creation. | 我们有离线的知识提取和创造, 知识图的创造。 |
| 201 | That whole part of memory is a big deal. | 这整个记忆的部分是一件大事。 |
| 202 | The other thing is, of course, tool use. | 另一件事当然是工具的使用。 |
| 203 | And we've taught these tools, these AIs, how to interact with the most important tools | 我们教过这些工具,这些AI,如何与最重要的工具互动 |
| 204 | that the humans ever made. | 是人类创造出来的 |
| 205 | It's no different than if you were a robot, it comes into our house, it's more likely | 和你是机器人一样,它进入我们家,更可能 |
| 206 | you're going to use all the tools that I have in the house than to go off and build it from scratch. | 你要用我家里所有的工具 而不是从头开始建造 |
| 207 | And so Spice is still going to be useful, and Verilog will be useful, and simulators | 所以香料还是有用的 维利洛格也是有用的 模拟器也是有用的 |
| 208 | 将是有用的, 和设计编译器, 和Photoshop, 和AutoCAD, 和细胞Katia, | |
| 209 | and all the Siemens tools. | 和所有西门子的工具。 |
| 210 | All of these different tools that we've been using to run our company, build our companies, | 所有这些不同的工具 我们一直在使用 管理我们的公司,建设我们的公司, |
| 211 | design things, operate things, they're all going to be great tools for the agents to use. | 设计一些东西,操作一些东西,它们都会成为特工们使用的伟大工具. |
| 212 | But the agents had to learn how to use it. | 但特工们必须学会如何使用它. |
| 213 | And so these two areas, knowledge and working memory and tool use, huge unlocks in the last year. | 因此,这两个领域, 知识和工作记忆以及工具的使用, 在去年,巨大的解锁。 |
| 214 | So now all of a sudden you could put these large language models encapsulated in their | 所以,现在突然你可以把这些大语言模型 封装在他们的 |
| 215 | harness or exoskeleton, and you can put them into your company, | 你可以把它们放进你的公司 |
| 216 | connect them to the information, the ground truth that you have, | 把它们与信息联系起来, 你拥有的真相, |
| 217 | connect them to the tools that they can use, and all of a sudden they start doing amazing things. | 把他们和他们可以使用的工具联系起来,突然他们开始做一些令人惊奇的事情. |
| 218 | Now the easiest part, frankly, was coding. | 坦白说,最简单的是编码 |
| 219 | That's the first on-ramp to useful work. | 这是第一个有用的工作。 |
| 220 | In a lot of ways, coding is easy because both the customer and the tools that they use | 在很多方面,编码是很容易的 因为客户和他们使用的工具 |
| 221 | are fairly accustomed to things that have tremendous use but are not extremely easy to use. | 相当习惯于具有巨大用途但并非极易使用的东西。 |
| 222 | In the early days of coding, it wasn't extremely easy to deploy, | 在编码初期,部署起来并不容易, |
| 223 | but software engineers found a way. | 但软件工程师找到了办法 |
| 224 | So now I think the next click for us, and you guys have probably been playing with Astra, | 所以,现在我想下一次点击我们, 你们可能一直在玩 Astra, |
| 225 | OpenAI Astra is really fantastic, and we're using it in all these different places. | OpenAI Astra真是太棒了, 我们正在所有这些不同的地方使用它。 |
| 226 | And so now the next phase of these agents, they're able to do design work and a lot of | 所以现在这些特工的下一阶段 他们能够做设计工作 |
| 227 | creative marketing work, and we're using it to go. | 创意营销工作,我们正在利用它去。 |
| 228 | You take a picture and give it to Astra. | 你拍张照给阿斯特拉 |
| 229 | It literally recreates it into CAD. | 它从字面上把它重现成CAD. |
| 230 | It recreates it into the CAD, which is incredible because we need the CAD so that we can interact | 它将它重现为CAD,这是不可思议的,因为我们需要CAD,这样我们才能互动 |
| 231 | with the CAD and make fine-tuning and adjust, and so Astra could really give us the first | 与CAD并做微调和调整, 所以Astra真的可以给我们第一个 |
| 232 | initial implementation. | 初步执行。 |
| 233 | Incredibly productive. | 令人难以置信的生产力。 |
| 234 | Now you take Astra and the latest release and you apply it to all of these different domains. | 现在,你把Astra和最新的发行版 应用到所有这些不同的领域。 |
| 235 | I think that besides software engineering, we're going to see designers and builders | 我认为除了软件工程之外,我们会看到设计师和建筑师 |
| 236 | and creators and this whole area of scientists and all of this whole unlocking. | 和创造者 和整个科学家领域 和所有这一切解锁。 |
| 237 | Now, the challenge, of course, is how do we deploy agents safely and securely into the | 现在,挑战当然是,我们如何安全、安全地部署特工人员 |
| 238 | world's enterprise? | 世界企业? |
| 239 | This is exactly the same question again with safety and security, and we'll talk about | 这个问题与安全保障完全相同 我们会谈谈 |
| 240 | that. | 那个 |
| 241 | We'll come in on that. | 我们进去吧 |
| 242 | What I'm struck by is that all of this occurred this year with OpenAI and Anthropic having, | 我感到惊讶的是,这一切发生在 今年的OpenAI和Anthropic有, |
| 243 | I don't know, two, three gigawatts of compute each. | 我不知道,两,三千兆瓦的计算。 |
| 244 | Relatively small amount of compute relative to what's coming. | 相对于即将到来的东西来说,计算数量相对较少. |
| 245 | Again, what I'm interested in is the way the economy generally works or people model the | 再说一遍,我感兴趣的是 经济的一般运作方式 或者人们模拟 |
| 246 | economy or they model a company is linear deceleration. | 经济或他们模拟公司 是线性减速。 |
| 247 | They're not very good at modeling acceleration, but here we have this double exponential. | 他们不太擅长建模加速, 但是我们这里有双倍指数。 |
| 248 | We have massive amount of new compute coming on board. | 我们有大量的新计算机上。 |
| 249 | So if OpenAI and Anthropic are, I don't know, three to five x-ing their compute next year, | 所以如果OpenAI和Anthropic是,我不知道, 三到五x - 在他们的计算明年, |
| 250 | does it stand to reason? | 这有道理吗? |
| 251 | We solved Navier-Stokes. | 我们解决了纳维尔 -斯托克斯 |
| 252 | We'll look back at this and say that was a tiny amount of compute and intelligence, | 我们回头看,说这是 微小的计算和智能, |
| 253 | the worst intelligence we're ever going to have from here forward solving a millennium | 我们最糟糕的智慧 从这里向前 解决一个千年 |
| 254 | problem. | 问题。 |
| 255 | So when you look at what's coming online next year, do you expect that we're going to have | 所以,当你看看明年上线时, 你期望我们将会有 |
| 256 | major breakthroughs in long running agentic problems that we can use these swarms of agents | 在长期运行的代理问题上的重大突破 我们可以利用这些毒剂群 |
| 257 | on? | 继续? |
| 258 | Well, Navier-Stokes is really a great example. | 纳维耶 -斯托克斯是一个很好的例子。 |
| 259 | Remember, OpenAI did not solve Navier-Stokes from nothing. | 记住,OpenAI没有从什么角度解决纳维耶-斯托克斯. |
| 260 | Correct. | 没错 |
| 261 | They solved it from the foundation of a lot of great scientists and a lot of great | 他们从许多伟大的科学家和伟大的科学家的 基础上解决了这个问题 |
| 262 | mathematicians and a lot of great previous work. | 数学家和很多伟大的前作 |
| 263 | Earlier I said we're at kind of the tipping point of a lot of these different fields | 之前我说过,我们正处在这些不同领域的关键点 |
| 264 | because we're kind of at that phase, that point with Navier-Stokes with a great deal | 因为我们在那个阶段 和纳维耶 -斯托克斯 有很多 |
| 265 | of foundational work already been done. | 基础工作已经完成。 |
| 266 | If we could just apply the exhaustive exploration of AI to the previous done work, we're going | 如果我们能把AI的详尽探索 应用到之前完成的工作,我们要去 |
| 267 | to see one breakthrough after another breakthrough rather quickly. | 很快地看到一个突破 一个又一个突破 |
| 268 | But we do need a lot of compute to do that. | 但我们确实需要大量的计算来做到这一点。 |
| 269 | And the reason for that is because a large amount of exploration was necessary. | 其原因是需要大量勘探。 |
| 270 | If it was simply somebody resolving some fundamental first principle theory in order | 如果只是某个人 解决一些基本的第一原则理论 |
| 271 | to discover the breakthrough, if it didn't require a vast amount of exploration of space, | 如果不需要大量探索空间 |
| 272 | search space, not space space, but search space, if it didn't require that, then somebody | 搜索空间,不是空间空间, 但搜索空间,如果它不需要, 那么有人 |
| 273 | would have found it earlier. | 早该发现的 |
| 274 | People are pretty smart. | 人们很聪明。 |
| 275 | So I think that there are many problems like this that are just on a precipice and really | 所以我认为,有很多这样的问题 只是处于一个边缘,真的 |
| 276 | wants to just have that massive amount of intelligence exploration to get us there. | 只是想进行大量的情报探索 才能把我们带到那里 |
| 277 | In the coming years, my expectation is that the cost of intelligence would fall fairly | 在未来的几年里,我期望情报费用会公平下降 |
| 278 | quickly. | 快点 |
| 279 | And the reason for that is because obviously we're bringing more compute online. | 其原因是,很明显,我们在网上带来了更多的计算。 |
| 280 | These models are getting smarter. | 这些模型越来越聪明了。 |
| 281 | When you're smarter, you could be more efficient. | 当你更聪明的时候,你会更有效率。 |
| 282 | And so the combination of these two things, the cost of intelligence will come down, | 所以这两件事的结合, 智能的成本会下降, |
| 283 | which is, of course, what they want. | 当然,这是他们想要的。 |
| 284 | And this is no different than Moore's Law. | 这跟摩尔定律没什么不同 |
| 285 | If you go back 30-some-odd years, we have three decades of computing cost decreasing by a factor | 如果你回到30多岁, 我们有30年的计算成本 下降因数 |
| 286 | of two every single year. | 每年两次。 |
| 287 | And we want intelligence costs to decrease by a factor of two or 20 every single year | 我们希望情报费用每年减少2或20倍 |
| 288 | because the consumption of it is going to more than make up for it. | 因为它的消耗 将超过弥补它。 |
| 289 | And so there's a massive empty space in the world that wants to be filled, | 世界上有一大片空地想被填满 |
| 290 | empty vacuum of opportunity, and we just got to get the cost down. | 空虚的机会真空, 我们只需要降低成本。 |
| 291 | I guess what stands out to me about Navier-Stokes is that they had 10,000 agents working together. | 我想纳维耶 -斯托克斯最突出的就是 他们有1万个特工合作 |
| 292 | To me, that sounds like a hyperscale company basically all pointing | 对我来说,这听起来像一个超规模的公司 基本上都指向 |
| 293 | agents to work together towards a goal. | 特务们一起为目标努力 |
| 294 | And I think in the early days right now, it's a lot of paralyzed search exploration. | 我想在早期,现在, 这是很多瘫痪的搜索探索。 |
| 295 | But I think the goal is that more of these agents can work together to a common goal | 但我认为目标就是 更多这些特工可以合作实现共同目标 |
| 296 | and that we can solve a lot more. | 我们可以解决更多的问题。 |
| 297 | And to your point, I think our math says it probably cost about $4 million of compute | 就你而言,我认为我们的数学说 可能要花费四百万的计算 |
| 298 | to solve this instantiation of Navier-Stokes. | 解决纳维耶-斯托克斯的这个即兴表演。 |
| 299 | The cost to do something like that over the next few years | 接下来几年做这种事的代价 |
| 300 | is going to drop by another 5x, 10x, right? | 还要再来个5x,10x,对吧? |
| 301 | So the number of things that we're going to be able to do is rapidly expanding. | 因此,我们将能够做的事情的数量迅速增加。 |
| 302 | Our ambition is going to go sky high. | 我们的雄心会飞升 |
| 303 | The original GPT probably cost, let's pick a number, say, maybe $20 million to do. | 最初的GPT可能要花费,让我们选一个数字,比如说,也许需要2000万美元做. |
| 304 | If you wanted to train GPT-1 now, it'll probably cost you, what, $1,000, $100 maybe? | 如果你现在想训练GPT-1, 可能要花你,什么,1000美元,100美元? |
| 305 | It's like $100, yeah. | 这就像100美元,是的。 |
| 306 | You went from $20 million, let's say $10 million, to $100, and we did that in 10 years. | 你从2000万美元,比如1000万美元,到100万美元, 我们在10年里做到了。 |
| 307 | And so I fully expect, meanwhile, our ambition got so large that the models that we're training now, | 所以我完全期望, 同时,我们的雄心壮大, 以至于我们现在训练的模型, |
| 308 | $3 trillion, $10 trillion, $30 trillion, we're talking about gigantic models now. | 3万亿,10万亿,30万亿 我们现在谈的是巨大的模型。 |
| 309 | Our ambition for the capabilities that technology has gone up, | 我们的雄心壮志 技术已经提升, |
| 310 | meanwhile, our use of it has gone up between 10 years ago to now, probably, I don't know, | 与此同时,我们使用它 在10年前到现在,也许,我不知道, |
| 311 | a million times. | 百万遍经. |
| 312 | And so I don't think this trend is going to stop. | 所以我不认为这种趋势会停止。 |
| 313 | There's a fairly insatiable demand for intelligence. | 相当无厌求智. |
| 314 | And the smarter you are, the more intelligent you are, the more ideas of applying intelligence | 你越聪明,你越聪明 运用智慧的想法越多 |
| 315 | you have. | 你曾经是 |
| 316 | The more I've learned in life, the more I realize there's more to do. | 生平所学愈多,觉悟愈多. |
| 317 | And that's because every time you learn something, it opens a new door. | 这是因为每次你学到一些东西, 它打开了一个新的门。 |
| 318 | Every time you learn something, it opens a new door. | 每次你学到一些东西, 它打开了一个新的门。 |
| 319 | And every time you open a new door, you see more opportunity, you see more challenges, | 每次你打开新门,你看到更多的机会,你看到更多的挑战, |
| 320 | you see more problems. | 你会看到更多的问题。 |
| 321 | And so that's the future. | 这就是未来 |
| 322 | People say, we're going to have more time. | 人们说,我们会有更多的时间。 |
| 323 | I completely disagree. | 我完全不同意这一点。 |
| 324 | I think for most people, this intelligence is going to unlock more opportunities. | 我想对大多数人来说,这种情报会释放更多的机会。 |
| 325 | And they're going to want to pursue it because they're ambitious. | 他们打算追求这个目标,因为他们雄心勃勃。 |
| 326 | Give me an example of a family, a father, a son. | 举一个家,一个父亲,一个儿子的例子. |
| 327 | A husband, a wife, a mother. | 丈夫,妻子,母亲 |
| 328 | Who doesn't want it? | 谁不欲. |
| 329 | Who don't have ambition for their families to have more? | 谁有抱负,欲家多得? |
| 330 | They just don't have the opportunity. | 他们只是没有机会。 |
| 331 | And these things are going to open up opportunities. | 这些东西将会打开机会。 |
| 332 | If you go back to 1930, Keynes talked. | 如果你回到1930年,凯恩斯说话。 |
| 333 | He made a prediction about the Industrial Revolution. | 他对工业革命作了预测. |
| 334 | And he said, in 100 years, the economy will have 8x'd, | 他说,100年后,经济将有8xd, |
| 335 | but the work week will go to 15 hours a week. | 每周工作15小时 |
| 336 | He was right. | 他说的对 |
| 337 | The economy did 8x, right? | 经济8X,对不对? |
| 338 | We had massive abundance and productivity gains from it. | 我们从中获得了大量的丰收和生产力的提高。 |
| 339 | Yeah, except we're working 80 hours a week. | 是啊,除了我们每周工作80小时。 |
| 340 | We didn't go to a 15-hour week. | 我们一周没去15小时 |
| 341 | And what he got wrong was the human spirit, right? | 他错了是人的精神,对吧? |
| 342 | He didn't realize that work is purpose. | 他没意识到工作是目的 |
| 343 | Discovery is purpose, right? | 发现是目的,对不对? |
| 344 | It's deeply encoded in, I think, our DNA. | 它深深地编码进了我们的DNA |
| 345 | And I'm working harder than I ever have, having more fun than I ever have. | 我比以前更努力工作 更快乐 |
| 346 | Not because I have to, like my grandparents did, right? | 不是因为我必须, 就像我的祖父母一样,对不对? |
| 347 | I'm doing it because I want to. | 我这样做是因为我想这样做。 |
| 348 | And so, I agree with you on that. | 因此,我同意你的观点。 |
| 349 | I'm fascinated with this idea. | 我对这个想法很着迷 |
| 350 | The singularity postulates that we've kind of now have all these agents that can work | 奇异的假设 我们现在有 这么多的特工可以工作 |
| 351 | together on long-running tasks. | 共同完成长期任务。 |
| 352 | They can stay on task. | 他们可以继续执行任务。 |
| 353 | I love your keep each other on task, working as a team. | 我喜欢你们互相配合 合作 |
| 354 | Hey, you're off. | 嘿,你离开。 |
| 355 | That's right. | 没错 |
| 356 | I love your idea that we build on top of all the things that have come, right? | 我喜欢你的想法,我们建筑 在所有的事物已经来了,对不对? |
| 357 | And so, if I think about full self-driving as an example, | 所以,如果我考虑 完全的自我驾驶作为一个例子, |
| 358 | we talked about it for a decade, and we really didn't get there, right? | 我们谈了十年,我们真的没有到那里,对不对? |
| 359 | But all of the sudden, like you go on Twitter, | 但突然间 就像你在推特上那样 |
| 360 | everybody experiencing Tesla, FSD, experiencing Waymo, we're here. | 每个人都在经历特斯拉,FSD, 经历Waymo,我们在这里。 |
| 361 | Somebody asked me, how was the drive here this morning? | 有人问我,今天早上的车怎么样? |
| 362 | I said, I don't know, because I got in my car, and I hit FSD, | 我说,我不知道, 因为我上了我的车, 我打了FSD, |
| 363 | and I didn't pay any attention until we pulled into the parking garage, right? | 直到我们拉进停车场,我才注意,对吧? |
| 364 | I was working the whole way here. | 我一路都在这里工作 |
| 365 | But if I think about that in the fields of science, | 但如果我在科学领域想到这一点 |
| 366 | the breakthroughs that are going to come in the fields of pharma and genetics, right? | 药物和遗传学领域的突破,对吧? |
| 367 | We've been laying down the tracks for 30 years. | 我们铺轨已经30年了 |
| 368 | We're going to break through on a lot of massive cures for the world. | 我们要突破 大量的大规模治愈世界。 |
| 369 | Of course, ChatGPT is already helping parents plan their day | 当然,ChatGPT已经 帮助父母计划自己的一天 |
| 370 | and solve their kids' health issue and helping them get better educated. | 并解决孩子的健康问题 帮助他们获得更好的教育 |
| 371 | Talk to us a little bit, though, as to why that's having such a hard time | 跟我们说说为什么这么难 |
| 372 | breaking through, Jensen. | 闯过去,詹森。 |
| 373 | We have all of these things that are out there improving lives | 我们有所有这些东西 在那里改善生活 |
| 374 | and that are right around the corner. | 就在拐角处 |
| 375 | And so, that's kind of curious to me as to why that's not breaking through. | 所以,我有点好奇 为什么它没有突破。 |
| 376 | I guess, let me take a shot at it. | 我想,让我试试看。 |
| 377 | Yeah. | 对 |
| 378 | And we'll ultimately not really know. | 我们最终不会真正知道。 |
| 379 | I've spoken to a lot of people. | 常与众说. |
| 380 | They told me what they thought. | 他们告诉我他们的想法 |
| 381 | I mean, one of the things, of course, is AI or super intelligence is soft matter. | 我的意思是,其中一件事,当然, 是AI或超级智能是软物质。 |
| 382 | You can't see it. | 你看不见它。 |
| 383 | And so, you've got the whole world talking about this technology, | 全世界都在谈论这个技术 |
| 384 | but you can't see it. | 但你看不见 |
| 385 | You can't touch it. | 不可触动. |
| 386 | And even though they're already benefiting from it, | 尽管他们已经从中获益了 |
| 387 | when they do search, they're already benefiting from it. | 当他们进行搜索, 他们已经从中获益。 |
| 388 | When they go shopping, they're benefiting from it. | 当他们去购物, 他们从中获益。 |
| 389 | When they're using Netflix, they're benefiting from it. | 当他们使用Netflix时,他们从中获益. |
| 390 | You go to Walmart, you're benefiting from it. | 你去沃尔玛,你从中获益。 |
| 391 | The fact that FedEx is getting stuff to us. | 联邦快递公司给我们带来麻烦 |
| 392 | It's all leveraging super intelligence and accelerated compute today. | 利用超级智能 加速计算 |
| 393 | But we don't do a very good job telling that story. | 但我们讲这个故事做得不好 |
| 394 | They're benefiting from it. | 他们从中获益。 |
| 395 | It's invisible. | 不可见者. |
| 396 | Right. | 对 |
| 397 | So, if we shut off the data centers and shut off super intelligence for a weekend, | 所以,如果我们关闭数据中心 关闭超级情报一个周末, |
| 398 | I think everybody would notice. | 我想大家都会注意到的 |
| 399 | I know you had a question about robotics in particular. | 我知道你有个关于机器人的问题 |
| 400 | I guess you notably said a few times this year that the chat GPT moment for robotics is here. | 我猜你今年已经说过几次了 机器人的聊天时间到了 |
| 401 | And I think you've also said that it could be your next $100 billion market. | 我想你也说过 这可能是你的下一个1000亿美元的市场。 |
| 402 | So, I guess, what are you seeing now in the inflection point in this year | 所以,我想,你现在看到什么 在今年的交际点 |
| 403 | that this is around the corner? | 这是在拐角处? |
| 404 | Yeah, I understated the $100 billion because we're already, | 是啊,我低估了1000亿美元 因为我们已经, |
| 405 | the first six months of this year, we did $10 billion in physical AI. | 今年头6个月,我们做了100亿美元 物理AI。 |
| 406 | Physical AI is basically AI that's somehow interacting with the physical world. | 物理AI基本上是与物理世界有某种相互作用的AI. |
| 407 | And it could be photon counting CT machines, it could be ultrasounds, | 可能是光子计算CT机 也可能是超声波 |
| 408 | it could be new imaging radars, next generation of 5G. | 可能是新的成像雷达 下一代5G |
| 409 | Not only does it understand radio waves, it's going to understand signals of all kinds. | 它不仅能理解无线电波,还会理解各种信号. |
| 410 | It's a sensing network. | 腾讯网. |
| 411 | Of course, robotics, self-driving cars, manufacturing systems, visual inspections, | 当然,机器人,自动驾驶汽车, 制造系统,视觉检查, |
| 412 | I mean, geospatial satellites, imaging the world and reconstructing it into 3D. | 我的意思是,地理空间卫星, 成像世界 重建成3D。 |
| 413 | So, there's basically, if you look at the entire electromagnetic spectrum, | 所以,基本上, 如果你看看整个电磁频谱, |
| 414 | there's AI stuff going on from visible to invisible spectrums. | 从可见光谱到隐形光谱 都有AI的东西 |
| 415 | And we're doing work across that entire thing. | 我们正在做整个事情。 |
| 416 | And as an AI, there's the perception part of it, there's the reasoning part of it, | 作为AI,有感知部分,有推理部分, |
| 417 | and then there's the planning and actuation part of it. | 然后是它的规划和激活部分。 |
| 418 | When you finally come to planning and actuation, then that's self-driving cars and robotics | 当你终于来到 规划和激活, 然后,这是自驾汽车和机器人 |
| 419 | and things like that. | 和类似的东西。 |
| 420 | In-warehouse delivery vehicles, pick and place robots that you don't have to program | 室内送货工具,选取和放置不需要编程的机器人 |
| 421 | as precisely because it learns by itself, all the way to people talking about consumer robots | 因为它自己学习, 一直到人们谈论消费机器人 |
| 422 | coming into your house, making your bed and things. | 进入你的房子, 做你的床和东西。 |
| 423 | Or healthcare robots, hospital robots, lots of really good examples there. | 或医疗机器人,医院机器人, 有很多很好的例子。 |
| 424 | Precisely delivering exactly your prescription to the patient, | 准确的将你的处方交给病人 |
| 425 | having robotic care, literally 24-7 every single second to keep patients safe. | 拥有机器人护理,从字面上说每秒24 -7 来保证病人的安全。 |
| 426 | And so, just lots and lots of great examples, | 所以,只是很多,很多伟大的例子, |
| 427 | entire surgical rooms that are going to be completely robotic. | 整个手术室都是机器人 |
| 428 | And so, lots of great technology being deployed. | 因此,许多伟大的技术被应用。 |
| 429 | And so, if you look at this whole thing, we're the world's largest physical AI company. | 所以,如果你看这整件事, 我们是世界上最大的物理AI公司。 |
| 430 | We did $10 billion in the first six months. | 我们在头六个月做了100亿美元。 |
| 431 | We did $10 billion in the last year. | 去年我们做了100亿美元。 |
| 432 | And that was our first $10 billion year. | 那是我们的第一个100亿年。 |
| 433 | The year before that, we did five. | 前一年,我们做了5次。 |
| 434 | And so, we went five, 10, let's say we did 20, who knows? | 所以,我们去了5,10, 让我们假设我们做了20, 谁知道呢? |
| 435 | So, maybe the hundred was a little understated. | 所以,也许这100 是一个有点低估。 |
| 436 | And I think the reasoning part of it is everything that matters to us in the final analysis | 我认为其中的推理部分 对我们来说是最重要的 |
| 437 | touches the physical world, and we live in the physical world. | 触摸物理世界, 我们生活在物理世界。 |
| 438 | And so, I think the market opportunity for bringing software, computing, intelligence | 所以,我认为市场的机会 带来了软件,计算,智能 |
| 439 | into an industry that largely historically was mechanical, electrical, that's a big opportunity. | 进入一个在历史上基本上是机械,电气的产业, 这是一个巨大的机遇. |
| 440 | I mean, look at obviously what Elon has done to the automotive industry, | 我的意思是,看看很明显 伊隆对汽车工业做了什么, |
| 441 | turned it into essentially a computing industry, a technology industry. | 把它变成一个基本上 计算工业,一个技术工业。 |
| 442 | And so, I think every car company is going to be a technology company. | 所以,我认为每个汽车公司 都会成为一个技术公司。 |
| 443 | Every shipping and receivings company, | 每个航运和接收公司, |
| 444 | logistics company is going to be a technology company. | 物流公司要成为科技公司. |
| 445 | I'm super excited for FedEx. | 我对联邦快递非常兴奋 |
| 446 | They're obviously going to be an AI company. | 他们显然会成为一个AI公司。 |
| 447 | They're obviously going to be a technology company. | 他们显然要成为一个技术公司。 |
| 448 | Walmart, obviously going to be a technology company, an AI company. | 沃尔玛,显然要成为一个技术公司,一个AI公司. |
| 449 | So, super excited for them. | 所以,超级兴奋他们。 |
| 450 | The entire telecommunications industry will be reinvented again into a technology industry. | 整个电信业将再次被重塑为技术产业. |
| 451 | They were a technology industry a long time ago, then became a services industry. | 早在很久以前,它们就是一个技术产业,然后成为了一个服务业。 |
| 452 | They're going to be a technology industry again. | 他们将再次成为科技产业. |
| 453 | And I'm delighted to work with AT&T and Verizon and Europe around the world | 我很高兴和AT&T、Verizon以及欧洲合作 |
| 454 | as they reinvent themselves. | 他们重新创造自己。 |
| 455 | And so, you take a look at every one of these industries that touch the physical world, | 所以,你看看每一个这些行业 触摸物理世界, |
| 456 | and there's a great idea for reinventing them with AI. | 并有一个伟大的想法,用AI来重塑它们. |
| 457 | You know, I was just in Washington. | 你知道,我只是在华盛顿。 |
| 458 | It's so weird. | 这太奇怪了。 |
| 459 | People put us in chips. | 人们把我们放在筹码里。 |
| 460 | And so, we're actually in warehouse and logistics. | 所以,我们实际上在仓库和物流。 |
| 461 | Yes. | 对 |
| 462 | We're in healthcare. | 我们在医疗领域 |
| 463 | We're in automotive. | 我们在汽车公司 |
| 464 | We're in telecom. | 我们在电信 |
| 465 | Every industry on the planet. | 世间一切业. |
| 466 | And so, we're in life sciences. | 因此,我们在生命科学领域。 |
| 467 | So, all these entertainment, obviously for a long time, | 因此,所有这些娱乐, 显然在很长一段时间, |
| 468 | NVIDIA was really vastly an entertainment company. | NVIDIA其实是一个巨大的娱乐公司。 |
| 469 | So, each one of these verticals, NVIDIA has a massive stack that's inside of it. | 因此,每个垂直的, NVIDIA 都有一个巨大的堆积 它的里面。 |
| 470 | And we're very deep in the ecosystem. | 我们非常深入的生态系统。 |
| 471 | And we have excellent partnerships there. | 我们在那里有很好的合作关系。 |
| 472 | And so, one of these days, people might not categorize us as a chip company anymore. | 因此,有一天,人们可能不再把我们归类为芯片公司。 |
| 473 | The playbook would be to take the Alpamio automotive, | 剧本是用阿尔帕米奥汽车 |
| 474 | I guess, playbook and insert it into each one of these industries. | 我猜,游戏本 并插入到每一个这些行业。 |
| 475 | Yeah, that's the final contribution. | 是啊,这是最后的贡献。 |
| 476 | But in each one of these, it takes three computers to build a physical AI platform. | 但在其中的每一个中,需要三台计算机来建立物理AI平台. |
| 477 | And so, you need a computer to create the brain. | 所以,你需要一台电脑来创造大脑。 |
| 478 | That's the training system. | 这就是训练系统。 |
| 479 | But very importantly, and this is where you're going to see the labs also do this, | 但非常重要的是,这就是你将要看到 实验室也这样做, |
| 480 | they need to build a digital twin of the internet | 他们需要建立一个数字双网络 |
| 481 | so that they don't have to put their AI agents out into the internet to learn how to use it. | 这样他们就不用把他们的AI特工 放到互联网上去学习如何使用 |
| 482 | And so, just like we created Omniverse as the digital twin | 所以,就像我们创造了Omniverse 作为数字双 |
| 483 | of warehouses and buildings and cities and the physical world, | 仓库、建筑物和城市以及有形世界, |
| 484 | we put our robot into the digital twin. | 我们把机器人放进数码双胞胎 |
| 485 | The digital twin learns how to be a robot. | 数码双胞胎学习如何成为机器人. |
| 486 | And then we deploy it into the last computer, which is the computer that's inside, | 然后我们把它放到最后一台电脑里, 也就是里面的电脑, |
| 487 | for example, a robot or a car or a surgical instrument or whatever it is. | 例如,机器人、汽车、手术器械等等。 |
| 488 | And so, it takes three robots, three computers and three complete stacks. | 因此,它需要三个机器人,三个计算机和三个完整的堆栈。 |
| 489 | And we build complete stacks for each one of them. | 我们为他们建造完整的堆栈。 |
| 490 | We're the only company in the world that does that. | 我们是世界上唯一这样做的公司 |
| 491 | And then, if you're a company that wants to build it end to end, | 然后,如果你是一个公司 想建造它结束到结束, |
| 492 | you could use two or three or one of our computers. | 你可以用我们两三台电脑 |
| 493 | We don't care. | 我们不在乎。 |
| 494 | And so, each one of a computer has its own software stack. | 因此,每个计算机都有自己的软件堆。 |
| 495 | We do the same across all of our verticals. | 我们的垂直也一样。 |
| 496 | But of course, these capabilities cut both ways, which is why | 但当然,这些能力 削减了两种方式, 这就是为什么 |
| 497 | we hear a lot of people talking about safety. | 我们听到很多人谈论安全 |
| 498 | You've been highly critical of the Doomer narrative. | 你一直高度批评杜默的叙述。 |
| 499 | You can only get there if you build the technology responsibly. | 只有负责任地建立技术,你才能到达那里。 |
| 500 | That's right. | 没错 |
| 501 | And so, the first thing is, build the technology responsibly. | 因此,第一件事就是, 负责任地建立技术。 |
| 502 | Notice the word build. | 注意这个词的构建。 |
| 503 | It's an engineering word. | 工程学词. |
| 504 | It's about technology. | 腾讯科技. |
| 505 | It's about engineering. | 论工学. |
| 506 | Engineers, scientists, researchers, we have to build this technology safely, responsibly. | 工程师,科学家,研究人员,我们必须安全,负责任地建立这种技术. |
| 507 | And the question is, can you? | 问题是,你能吗? |
| 508 | And the answer is absolutely yes. | 答案是绝对是的。 |
| 509 | Because if you can't, then you shouldn't build it at all. | 因为如果你不能,那么你不应该建造它。 |
| 510 | Of course. | 当然 |
| 511 | And so, we either have to believe we can or we shouldn't build it. | 所以,我们要么相信 我们可以或不应该建造它。 |
| 512 | Okay? | 好吗? |
| 513 | And so, I believe we can. | 所以,我相信我们可以。 |
| 514 | And so, let me just reason about it with you. | 让我跟你解释一下 |
| 515 | So, the first part of it is, some people say that AI is uncontrollable because it's impossible | 所以,第一部分是,有些人说AI是不可控制的,因为它是不可能的 |
| 516 | to align perfectly. | 完全对齐。 |
| 517 | Alignment means this for all the audience. | 对齐意味着所有观众都这样 |
| 518 | You tell somebody, I want you to do this. | 你告诉别人,我要你这样做。 |
| 519 | And so, the mission is clear. | 因此,任务是明确的。 |
| 520 | The mission is clear. | 任务明确. |
| 521 | And it's clear to you because you said it in English. | 而且你很清楚 因为你用英语说 |
| 522 | And so, it could be, I want you to solve this quadratic equation. | 所以,这可能是, 我希望你能解决这个四极方程。 |
| 523 | And so, the mission is to solve the quadratic equation. | 因此,任务是解决四极方程. |
| 524 | The alignment is, how do you do it? | 校对是,你怎么做到的? |
| 525 | There are many ways to do it. | 有多种方法可以做到. |
| 526 | For example, you memorized it. | 例如,你记住了。 |
| 527 | And so, the answer is, whatever it is. | 所以,答案是,无论它是什么。 |
| 528 | Or, you saw the answer in a textbook, so you looked it up. | 或者,你看到答案 在教科书, 所以你查了一下。 |
| 529 | You refer to a reference. | 你指的是一个参考文献。 |
| 530 | Or, you went to a drawer and you know the drawer had all the answers. | 或者你去抽屉 你知道抽屉里有答案 |
| 531 | You opened the drawer, you picked out the answer. | 你打开抽屉,你挑出答案。 |
| 532 | Some people call that cheating. | 有人称之为作弊 |
| 533 | I call that opening the drawer and retrieving the answer. | 我叫它打开抽屉 找回答案 |
| 534 | And that's what computers do. | 电脑就是这样的。 |
| 535 | There's a lot of different ways you could solve that problem. | 有很多不同的方法 你可以解决这个问题。 |
| 536 | Now, of course, that one is very explicit. | 现在,当然,这个非常明确。 |
| 537 | Here's another one. | 这里还有一个。 |
| 538 | Be successful in life. | 成就众生. |
| 539 | The mission is easy to say. | 任务很简单 |
| 540 | And in fact, we say it to each other all the time. | 事实上,我们经常互相说 |
| 541 | We say to our children, we say to our friends, | 我们对我们的孩子,我们对我们的朋友, |
| 542 | let's do our best to contribute to society. | 让我们尽力为社会作出贡献。 |
| 543 | Well, now, the question is how? | 那么,现在,问题是如何? |
| 544 | There are so many ways. | 有这么多的方法。 |
| 545 | How do you reasonably align in everybody's culture and background? | 你如何合理调整每个人的文化和背景? |
| 546 | Some people think that contributing to society means to be generous, to be kind. | 有些人认为对社会的贡献意味着慷慨,善良. |
| 547 | Some people might mean that contributing to society | 有些人可能意味着对社会的贡献 |
| 548 | make a lot of money and then be philanthropic. | 赚了很多钱,然后是慈善的。 |
| 549 | Some people might, so on and so forth. | 一些人可能,比如说。 |
| 550 | Isn't that right? | 是吗? |
| 551 | Yes. | 对 |
| 552 | Okay. | 摆 |
| 553 | And some people might just think that building a company alone | 一些人可能只是认为 建立一个公司单独 |
| 554 | or doing something alone is a sufficient contribution to society. | 或单独做一些事情,是对社会的充分贡献。 |
| 555 | Creating a beautiful piece of art or | 创造美丽的艺术品,或者 |
| 556 | all of it is true. | 这一切都是真的。 |
| 557 | And so the question is, what's the alignment? | 所以问题是 校对是什么? |
| 558 | And it just depends on your background, your culture, | 这取决于你的背景,你的文化, |
| 559 | maybe has something to do with where you live, where you came from. | 可能和你住在哪里,你来自哪里有关 |
| 560 | There's just an infinite number of ways. | 唯有无数法. |
| 561 | And so my point is this, alignment is going to be a forever problem. | 所以我要说的是 校正将永远是一个问题 |
| 562 | If we depend on perfect alignment in order to ensure safety, | 如果我们依靠完美的配合来确保安全 |
| 563 | then we're never going to get there. | 那我们就永远去不了了 |
| 564 | And so the answer is this. | 所以答案是这样的。 |
| 565 | The way we treat, the way we think about babies | 我们对待孩子的方式 |
| 566 | and when we were raising kids and we have pets or whatever it is, employees. | 当我们抚养孩子的时候 我们有宠物或什么的 员工 |
| 567 | You start with containment. | 你从隔离开始 |
| 568 | You give them the rights, the access to tools, the access to information, | 你给他们权利 获得工具 获得信息 |
| 569 | the access to whatever responsibility, capabilities that they have, funds, resources. | 获得其拥有的任何责任、能力、资金和资源。 |
| 570 | You give them access based on the job that you ask them to do. | 你根据你要求他们做的工作 给他们机会 |
| 571 | You start out with zero trust. | 你从零信任开始 |
| 572 | When you hire somebody, you start out with zero trust. | 当你雇人时 你从零信任开始 |
| 573 | When you have a baby that's one year old, you put them in a playpen. | 当你有一个一岁大的婴儿, 你把他们在一个游戏机。 |
| 574 | You assume zero trust. | 你假设没有信任。 |
| 575 | And so zero trust, you give them minimum rights | 如此的零信任, 你给他们最低的权利 |
| 576 | and you expand that boundary as you learn more. | 当你学到更多东西时,你就会扩大边界。 |
| 577 | And so whenever we have a new AI, we have to assume the same thing. | 所以,每当我们有了新的AI, 我们必须假设同样的事情。 |
| 578 | Whenever you're testing an AI, zero trust, minimum boundaries, minimum rights, | 当你测试一个AI, 零信任,最小界限,最小的权利, |
| 579 | and you expand it as you test and give yourself confidence. | 在测试时扩展它 并给自己信心 |
| 580 | So containment, very tight containment is really important to any autonomous system. | 所以控制,非常严密的控制 对任何自主系统都很重要。 |
| 581 | Second thing you do, monitoring. | 第二件事你做,监视。 |
| 582 | And the monitoring has to be out of band. | 而监控必须脱离波段. |
| 583 | And the reason for that is because if you tell an intelligent person, | 原因是如果你告诉一个聪明人 |
| 584 | hey, listen, I'm going to sit here and watch you do your work to make sure you're doing it right. | 嘿,听着,我要坐在这里 看着你做你的工作 确保你这样做是对的。 |
| 585 | I'm going to sit here and watch you make sure that you're serving customers properly | 我要坐在这里看着你 确保你为顾客服务 |
| 586 | and that you're working hard enough. | 而且你已经够努力了 |
| 587 | That autonomous system is going to behave differently than when you're not watching. | 这个自主系统的行为会和你不看的时候不同 |
| 588 | It's like Brad's in the room with me or not. | 就像布拉德在我房间里一样 |
| 589 | Exactly, exactly. | 没错,没错。 |
| 590 | And so the monitoring has to be out of band. | 因此,监测必须脱离范围。 |
| 591 | And it has to have the ability to observe that the autonomous system is doing something that | 它必须有能力观察 自主系统正在做一些事情 |
| 592 | was inconsistent with the policy or inconsistent with alignment. | 与政策不一致或与协调不一致。 |
| 593 | And you have the ability to either escalate or shut it down, just turn it off. | 你有能力或升级 或关闭它, 只是关闭它。 |
| 594 | And so those kind of capabilities have to exist. | 所以这种能力必须存在。 |
| 595 | That's engineering, that's technology, and it's possible. | 这是工程技术,这是可能的。 |
| 596 | 我们最近宣布它叫OpenShelf。 | |
| 597 | We have over 100 companies now joining us. | 现在有100多家公司加入我们。 |
| 598 | It's got lots and lots of activity around the world because everybody wants the same thing. | 它在世界各地有很多活动,因为每个人都想要同样的东西. |
| 599 | Starting at the highest level, when I looked around the table at the White House, | 从最高层开始 当我在白宫的桌子周围看时 |
| 600 | everybody around that table, their parents, they have kids. | 桌旁的每一个人, 他们的父母,他们有孩子。 |
| 601 | I think there's this view in the country that there are some leaders of super intelligence | 我觉得这个国家有种观点认为 有一些超级情报的领袖 |
| 602 | that are just yoloing, that don't really care, they're being reckless, | 那只是优柔寡断, 并不在乎, 他们正在鲁莽, |
| 603 | they just want to grab all the money. | 他们只是想抢走所有的钱。 |
| 604 | And yet when I looked around that table, there wasn't a person in there that I saw | 然而当我在桌子周围看时, 里面没有我看见的人 |
| 605 | that wants to be reckless, that wants to yolo. | 想要鲁莽 想要尤洛 |
| 606 | In fact, everybody there wanted to find the right balance between fast and safe. | 事实上,那里的每个人都想在快速和安全之间找到正确的平衡。 |
| 607 | That was the president's directive. | 这是总统的指示。 |
| 608 | And we debated a little bit what that meant. | 我们辩论了一下这意味着什么。 |
| 609 | There are different versions of how to implement that. | 关于如何执行这一点,有不同的版本。 |
| 610 | But what would you say to the American public who fears that we have a reckless president | 但是,你会怎么对美国公众说 谁担心我们有一个鲁莽的总统 |
| 611 | or reckless leaders? | 还是鲁莽的领导人? |
| 612 | Did you see any evidence of that? | 你看见什么证据了吗? |
| 613 | You have kids. | 你有孩子吗? |
| 614 | Do you worry about the future that you're bringing into existence with super intelligence? | 你担心你用超智能带来的未来吗? |
| 615 | Safety and security, technology are part of AI. | 安全保障,技术是AI的一部分. |
| 616 | There's some weird formulation that AI, technology, and safety and security are orthogonal. | 有一种怪异的表述认为AI,技术,安全保障是正统的. |
| 617 | It's like the airbags and the braking system and the radars, the cameras, | 就像气囊 制动系统 雷达 摄像头 |
| 618 | the ABS, the lane keeping, all of that technology is somehow independent of the car. | ABS,车道的维护, 所有的技术 在某种程度上是独立于汽车。 |
| 619 | Do you guys understand what I'm saying? | 你们明白我在说什么吗? |
| 620 | The only thing that matters to the car is the engine. | 汽车唯一要紧的是引擎. |
| 621 | In fact, the whole car is part of the car's technology. | 实际上,整车是汽车技术的一部分. |
| 622 | When I say we need to accelerate AI's technology, what I mean by that | 当我说我们需要加速AI的技术,我的意思是 |
| 623 | is we need to accelerate the security technology. | 我们需要加快安全技术。 |
| 624 | We have to accelerate the safety technology. | 要加快安全技术建设. |
| 625 | They all have to be accelerated. | 它们都必须加快速度。 |
| 626 | If you ask me right now, I think everybody would say, | 如果你现在问我, 我想每个人都会说, |
| 627 | let's accelerate security. | 让我们加快安全。 |
| 628 | Let's accelerate safety technology. | 加快安全技术建设. |
| 629 | Nobody says, no, let's pace that. | 没有人说,不,让我们的速度。 |
| 630 | We should accelerate that. | 我们应该加快步伐。 |
| 631 | It's all part of the same thing. | 皆是同义. |
| 632 | You can't separate capability, functionality from safety and security. | 无法将能力,功能与安全保障分开. |
| 633 | They go us one. | 他们去我们一个。 |
| 634 | When you deliver the service, nobody's going to say, | 当你交付服务, 没有人会说, |
| 635 | I really like the functionality of your service, but boy, is it wrecking havoc. | 我真的很喜欢你的服务功能, 但是男孩,它是破坏破坏。 |
| 636 | I think it's all one thing. | 我认为这都是一件事。 |
| 637 | We treat it as all one thing, and the industry has to treat it as one thing. | 我们把它当作是一回事,工业必须把它当作一件事。 |
| 638 | When we say accelerate, we have to mean accelerating the entire package. | 当我们说加速时,我们必须指加速整个一揽子计划。 |
| 639 | Right now- | 现在... |
| 640 | Including the safety. | 包括保险箱 |
| 641 | Including, especially the safety. | 包括,特别是安全。 |
| 642 | You mentioned pacing. | 你提过节奏 |
| 643 | Especially safety. | 特别是安全 |
| 644 | You know this. | 你知道这一点。 |
| 645 | Some people read Pacing the Frontier by Dario and thought that meant slowing down. | 有些人读达里奥的"边疆之平",认为这意味着减速. |
| 646 | I thought a funny moment in the room was when the president called on his new best friend Dario. | 我想在房间里一个有趣的时刻 是当总统召唤 他的新最好的朋友达里奥。 |
| 647 | He referred to him in that way, and he said, explain to me your version of fast and safe. | 他用这种方式提到他,他说, 向我解释你的观点 快速和安全。 |
| 648 | Dario started off by saying, Mr. President, you said whoever wins AI wins. | 达里奥一开始说,总统先生 你说谁赢了AI谁赢了 |
| 649 | He said that we're all speeding up, that they're speeding up, that they want to go fast, | 他说,我们都在加速, 他们正在加速, 他们想要去快, |
| 650 | that they want to win the race with China, but also acknowledge that along the way, | 他们想赢得与中国的比赛,但也承认 一路, |
| 651 | to your point you just made, we have to make sure the safety systems are also speeding up | 你刚才说的,我们得确保安全系统也加速 |
| 652 | and also keeping pace. | 还要跟上步伐 |
| 653 | I thought that was an important moment because I think people have positioned him as though | 我认为这是一个重要的时刻 因为我认为人们已经定位他 |
| 654 | he's in opposition to this, but it seemed to me that you were all very much on the same page. | 他反对这个 但在我看来 你们都一样 |
| 655 | You went to the cabinet room with him and Greg and the other frontier labs to sign | 你和他、格雷格 以及其他前沿实验室去柜房签字 |
| 656 | this agreement. | 这个协议。 |
| 657 | Was there any chatter in that room about differences of opinion or was everybody, | 那间屋子里有没有关于意见分歧的闲聊 或者是每个人 |
| 658 | you obviously got it signed, did anybody have to be put in a headlock or | 你显然得到了它签署, 有没有任何人必须 被置于一个头锁或 |
| 659 | did everybody go willingly? | 大家都是自愿的吗? |
| 660 | No, everybody was very constructive, very productive. | 不,每个人都很有建设性,很有成效。 |
| 661 | I think that without exception, everybody is dedicated to advancing the industry safely | 我认为,无一例外,每个人都致力于安全推进这个行业。 |
| 662 | and securely. | 安全无虞 |
| 663 | And the reason for that is because if it's not safe, it's not secure, the functionality | 原因是如果它不安全, 它不安全,功能 |
| 664 | will never get a chance to speak, and so the capability will never get a chance to shine. | 永远不会有说话的机会, 所以能力永远不会有发光的机会。 |
| 665 | And so we know that the priorities are at all times safety and security. | 因此我们知道,优先事项在任何时候都是安全和保障。 |
| 666 | Of course, recognizing that this industry came through a transition just recently. | 当然,认识到这个行业刚刚经历了一个转变。 |
| 667 | It wasn't but six months ago, this technology barely worked. | 6个月前,这个技术几乎没用 |
| 668 | Without functionality and capability, what's the point in making it safe and secure? | 没有功能和能力,何必让它安全无虞? |
| 669 | Nobody even cares about it. | 没人在乎 |
| 670 | Obviously, the first job of AI is to demonstrate it could do something useful, | 显然,AI的第一项工作就是证明它可以做一些有用的事情, |
| 671 | and it really reached that point six months ago. | 六个月前就已经到了这个地步了 |
| 672 | And so now, it really wants to be deployed and broadly deployed. | 因此,现在,它真正想要部署和广泛部署。 |
| 673 | And so there's no question that the entire industry now has a strategic imperative, | 所以毫无疑问,整个行业 现在有战略需要, |
| 674 | an incredible imperative, and great urgency to make sure that whatever we deploy | 不可思議的迫切性 确保无论我们部署什么 |
| 675 | is safe and secure. | 安全无虞 |
| 676 | So I think everybody is completely convinced, aligned, functionality, capability, safety, | 所以我认为每个人都完全相信, 一致,功能,能力,安全, |
| 677 | security, they're all at exactly the same place, priority number one. | 保安,他们都在同一地点,第一优先 |
| 678 | And so the question is winning. | 所以问题就是赢 |
| 679 | I would reframe winning not of the technology industry winning, | 我将重新塑造赢得 而不是技术行业赢得, |
| 680 | not of just the AI industry winning. | 不只是AI行业赢了。 |
| 681 | America winning means every industry in America winning. | 美国获胜意味着美国每一个行业都获胜. |
| 682 | Health care. | 保健。 |
| 683 | That's right. | 没错 |
| 684 | Nothing would give me more joy. | 没有比这更让我高兴的了 |
| 685 | The definition of America winning is when we talk about Lilly winning, | 美国获胜的定义是 当我们谈论莉莉获胜, |
| 686 | Walmart winning, FedEx winning, every single one, Boeing winning, | 沃尔玛获胜 联邦快递获胜 每个人 波音获胜 |
| 687 | the US military winning. | 美国军方获胜。 |
| 688 | Every citizen winning. | 每个市民都会赢 |
| 689 | And so our university is winning. | 所以,我们的大学正在赢得。 |
| 690 | Our high school is winning. | 我们的高中是胜利。 |
| 691 | That's the definition of winning. | 这就是获胜的定义. |
| 692 | There is no winning if it's just the technology won, but nobody used it. | 如果只是技术赢了,就没有胜利,但没人用过它。 |
| 693 | The greatest risk is, of course, I'll just say baseline. | 最大的风险当然是基准 |
| 694 | There will be no AI technology industry if we don't build it safely and securely. | 如果我们不安全可靠地建造AI技术产业,就不会有. |
| 695 | That this is an existential issue for us. | 对我们来说,这是一个存在的问题。 |
| 696 | And if we make it an existential industry for everybody else, | 如果我们让它成为其他人的生存产业 |
| 697 | well, that's not going to happen because it would be an existential issue for us first. | 好吧,那不会发生 因为它会是一个存在的问题 首先对我们来说。 |
| 698 | So having said that, the definition of winning for us, | 所以说,我们赢了的定义, |
| 699 | and this is the part I always have to change a little bit. | 这就是我总得改变的部分 |
| 700 | I want us to pivot. | 我想我们做点什么 |
| 701 | I like that. | 我喜欢 |
| 702 | I want us to pivot. | 我想我们做点什么 |
| 703 | And when I see AI winning, America AI winning, | 当我看到AI赢了 美国AI赢了 |
| 704 | I'm looking down from these days, we spent a lot of time in the air. | 我从这些天开始往下看 我们花了很多时间在空中 |
| 705 | And we're looking down. | 我们正往下看 |
| 706 | Every state's winning. | 每个州胜. |
| 707 | Every county's winning. | 各县胜. |
| 708 | Every community is winning. | 每个社区都在赢得胜利。 |
| 709 | Every company's winning. | 每个公司都赢 |
| 710 | Everybody won because we could afford to spend more on health care. | 每个人都赢了,因为我们可以负担更多的医疗开支。 |
| 711 | We could afford to spend more on education. | 我们可以负担更多的教育开支。 |
| 712 | We could afford to spend more on interstate commerce, highways, etc. | 我们可以在州际商业,高速公路等方面多花钱 |
| 713 | All of the infrastructure of the country. | 国家所有基础设施。 |
| 714 | What I hear you saying is the definition of winning is that when we look back 25, | 我听到你说的胜利的定义是 当我们回顾25, |
| 715 | 30, 40 years from now, the economy needs to be 10 times as large. | 30,40年后经济规模要达到10倍. |
| 716 | Everybody's lives need to be better. | 每个人都需要更好的生活。 |
| 717 | And existential and primary to all of that is it has to be done in a way that's aligned | 而存在和首要的 这一切是 它必须以一种一致的方式进行 |
| 718 | with human interest, which this is the last question I have on this. | 与人类的兴趣, 这是我对这个最后一个问题。 |
| 719 | And I know we have a lot to get to. | 我知道我们有很多事情要做 |
| 720 | Most people don't understand what our security framework is in the country. | 大多数人不明白 我们的安全框架在这个国家。 |
| 721 | I've said we have the most robust superintelligent security framework in the world. | 我说过我们有世界上最强大的超级智能安全框架 |
| 722 | We're clearly the only country on the planet that's actually sanctioned a model | 我们显然是地球上唯一一个 真正批准模型的国家 |
| 723 | and pulled it off the shelf with mythos earlier this year. | 今年早些时候用神话把它从架子上拉下来 |
| 724 | But I describe it as a three-leg stool. | 但我形容它是一个三脚凳。 |
| 725 | We have the Frontier Model Review of every new frontier model by Casey | 我们有凯西对每一个新疆域模型的前沿模型审查 |
| 726 | and the Department of Commerce. | 和商务部。 |
| 727 | So they do all their work. | 所以他们做所有的工作。 |
| 728 | If a new model is coming out, they review it. | 如果一个新的模型出现, 他们审查它。 |
| 729 | They test it for whether or not they think it's going to cause a problem in the marketplace. | 他们测试它是否认为它会在市场上引起问题。 |
| 730 | Now we have the White House Accord on Superintelligence that you helped to architect. | 现在我们有了你帮助建筑师的 《白宫超级情报协议》 |
| 731 | Internal and external audits, board oversight, and all of the liability that comes with that. | 内部和外部审计、董事会监督以及随之而来的所有责任。 |
| 732 | And then third, we have the SI Force, now led by Jay Clayton, of course, | 然后第三,我们得到了SI Force, 现在由杰·克莱顿领导,当然, |
| 733 | Emil and Scott and Andrew. | 埃米尔和斯科特和安德鲁。 |
| 734 | And that's doing critical industry coordination, cyber, et cetera. | 这正在做关键的行业协调,网络等等。 |
| 735 | So these three legs on the stool, there are people out there, | 凳子上的这三条腿 外面有人 |
| 736 | I heard Gates recently saying we're not doing enough. | 我最近听到盖茨说我们做得不够 |
| 737 | We need new laws on the books and everything else. | 我们需要新的法律 关于书籍和其他一切。 |
| 738 | That three-legged stool seems like a lot to me in a very short period of time | 那个三脚凳在很短的时间内 对我来说是很多 |
| 739 | that's come together as a leader of one of the most important companies, | 作为最重要的公司之一的领导者, |
| 740 | obviously, in superintelligence. | 显然,在超级情报。 |
| 741 | Do you feel comfortable for that future that we're trying to usher in? | 你对我们要迎接的未来感到舒服吗? |
| 742 | That is a good step forward where we need to be right now. | 这是向前迈出的好一步,我们现在需要这样做。 |
| 743 | Yeah, the best way to get to the future is take action quickly. | 是啊,最好的办法到达未来 是迅速采取行动。 |
| 744 | And if we argue about what's perfect versus what's first, | 如果我们争论什么是完美的,什么是第一, |
| 745 | we'll end up arguing about perfection and we'll never get there. | 我们最终会争吵 关于完美的,我们永远不会到达那里。 |
| 746 | And so I love the fact that these three-legged stool are actionable, | 所以我喜欢这个三脚凳可以行动的事实 |
| 747 | they're tangible, they have real teeth. | 它们是有形的, 他们有真正的牙齿。 |
| 748 | Right. | 对 |
| 749 | It's happening now. | 现在已经发生了 |
| 750 | It's happening right now. | 正在发生 |
| 751 | We're doing it now. | 我们正在做。 |
| 752 | We're taking action now. | 我们正在采取行动。 |
| 753 | And the reason we're taking action now is because it's good for our company, | 而我们现在之所以采取行动 是因为这对我们公司有好处 |
| 754 | it's good for our industry, it's good for our country, | 这对我们的工业有好处,对我国有好处, |
| 755 | it's good for our society, it's good all the way around. | 这对我们的社会有好处,对各地都有好处. |
| 756 | And if we don't do these things, the outcome is less desirable. | 如果我们不这样做,结果就不那么可取了。 |
| 757 | We don't get to pursue that optimistic future that we all dreamt about. | 我们无法追求我们都梦想中的乐观未来。 |
| 758 | And so if you go back and look at another industry, | 所以如果你回去看看另一个行业 |
| 759 | I don't remember another industry that is currently at its current state | 我不记得其他的行业 现在处于现状 |
| 760 | and have now formulated all of these cross-company, | 并且结识了所有这些跨公司, |
| 761 | industrial government to industry coordination to make sure that we | 工业政府与工业协调,以确保我们 |
| 762 | advance this in a responsible and optimistic way. | 以负责和乐观的方式推进这项工作。 |
| 763 | I just think it's so important for people to understand, | 我只是觉得这对人们来说很重要 |
| 764 | because when I hear senators talking about it on television, | 因为当我听到参议员在电视上谈论, |
| 765 | on both sides, to be perfectly honest. | 在两边,完全诚实。 |
| 766 | They have to learn about what we're doing here first. | 他们必须先了解我们在做什么 |
| 767 | They don't even know what we're doing. | 他们甚至不知道我们在做什么。 |
| 768 | We are doing an extraordinary amount. | 我们正在做一个非凡的金额。 |
| 769 | And by the way, these aren't voluntary things. | 顺便说一句,这些不是自愿的。 |
| 770 | There's not going to be a frontier model that gets released | 不会有边疆模型被放出来 |
| 771 | that Casey is not reviewing and the Commerce Department is not reviewing. | 凯西没有审查 商务部也没有审查 |
| 772 | And they've demonstrated they have the authority | 他们证明他们有权力 |
| 773 | under export controls to pull them off the shelf. | 在出口管制下把他们从货架上拉下来 |
| 774 | Do two things. | 做两件事 |
| 775 | One, come and spend some time and learn what actually has to be done. | 第一,来花点时间 了解实际需要做的事 |
| 776 | What's the problem? | 有什么问题吗? |
| 777 | If you want a solution, you've got to at least understand what the problem is. | 如果你想找到解决办法,至少你必须明白问题所在。 |
| 778 | So one, come and learn about it. | 所以第一,来学习一下。 |
| 779 | We've got plenty of time. | 我们有足够的时间。 |
| 780 | I've got all the time in the world for this kind of stuff. | 我为这种事有时间 |
| 781 | And just come talk to us, come talk to our industry and learn. | 来和我们谈谈, 来和我们的行业谈谈和学习。 |
| 782 | The second thing is learn about the solution that we put in place | 第二点是了解我们提出的解决方案 |
| 783 | and try to see if we can make that, if there are ideas to make that better. | 试着看看我们是否能够做到这一点, 如果有想法使这一点更好。 |
| 784 | Of course, we're all completely open-minded. | 当然,我们都完全开明。 |
| 785 | But be constructive. | 但要具有建设性。 |
| 786 | Right. | 对 |
| 787 | And I thought that was part of the fourth point on the accord on super intelligence | 我以为那是关于超级情报协议的第4点 |
| 788 | was this is not static. | 这不是静态的。 |
| 789 | This is dynamic. | 这是动态的。 |
| 790 | This group is going to get together routinely. | 这个小组将定期开会。 |
| 791 | We're going to update the framework as required when the technology evolves. | 科技发展时,我们将根据需要更新框架. |
| 792 | But I want to move forward now to compute. | 但我现在想继续计算 |
| 793 | Of course, long-running agents that work in teams of thousands or millions to solve | 当然 长期在千千万万团队工作的特工 |
| 794 | complicated tasks, it all consumes massively more compute. | 复杂的任务,都消耗了 大量的计算。 |
| 795 | Dylan, our friend at Semi Analysis, has said we're going to bring on board 43 gigawatts | 迪伦,我们在半分析公司的朋友, 已经说我们要带上43千兆瓦 |
| 796 | of new AI compute next year. | 新AI计算明年。 |
| 797 | Just to contextualize it, we'll share a chart with people online. | 为了了解情况,我们将在网上与人们分享一张图表。 |
| 798 | That's more than all of the AI compute that exists in the world today. | 这比今天世界上所有的AI计算要多. |
| 799 | So we're going to bring that on board next year. | 所以明年我们就要把这事办好 |
| 800 | The year after that, another 40 gigawatts of compute. | 之后的一年,再计算40千兆瓦. |
| 801 | You know where every single GPU is allocated. | 你知道每个GPU分配的地方。 |
| 802 | So as you sit here, does that sound right to you? | 所以当你坐在这里, 这听起来对吗? |
| 803 | Do you think we're going to bring on that much compute? | 你觉得我们会用这个计算吗? |
| 804 | And do you think that some of the things we're hearing out there about permitting issues | 你觉得我们听到的一些关于 允许的问题 |
| 805 | or bottlenecks in the supply chain or local protests, | 或供应链中的瓶颈或地方抗议, |
| 806 | do you think the number is going to be materially different than that based upon what you're seeing? | 你觉得根据你所看到的 数字会有实质性的不同吗? |
| 807 | I think the demand is definitely there and more. | 我认为需求肯定在那里,而且更多。 |
| 808 | The demand is definitely there and more. | 需求肯定在增加。 |
| 809 | The challenge, of course, is that there are a lot of obstacles. | 当然,挑战是有许多障碍。 |
| 810 | When you move from digital to the physical, there are a lot of obstacles. | 当你从数字移动到物理, 有很多障碍。 |
| 811 | There's land permitting. | 有土地允许。 |
| 812 | There's power generation that you've got to bring online. | 有发电,你必须带上网。 |
| 813 | Of course, you need construction workers and outfitters and plumbers and all kinds of— | 当然,你需要建筑工人 装配工和水管工 还有各种... |
| 814 | Created a million new jobs just building these things. | 创造了一百万个新工作 仅仅建造这些东西。 |
| 815 | Created a million new jobs and their skills are not fungible. | 创造了一百万个新工作,他们的技能是不可替代的。 |
| 816 | They have a lot of skills, and particularly the ones that are building these | 他们有很多技能 尤其是那些正在建造这些 |
| 817 | $50 billion data centers, you're going to want some people with great skills. | 五百亿的数据中心, 你会想要一些有优秀技能的人。 |
| 818 | There's a critical part of every single one of the build out of these things where | 每一个建筑中都有关键的一部分 |
| 819 | you need some very specific skills. | 你需要一些非常特别的技能。 |
| 820 | I think that—and then there's, of course, whatever moratoriums that are happening, | 我想,然后,当然,不管 暂停发生什么, |
| 821 | whatever issues that we're seeing around the country. | 无论我们在全国看到什么问题 |
| 822 | My guess is that land power shell, the build out resources will be one of the most important | 我猜,土地动力壳, 建造资源将是最重要的之一 |
| 823 | obstacles to reaching that. | 阻碍实现这一点。 |
| 824 | And so our company, as you know, now speaking specifically for us, | 所以,我们公司,你知道, 现在特别为我们说话, |
| 825 | there are several things that we're doing to mitigate that. | 我们正做一些事情来缓解这种情况。 |
| 826 | Because we're the only company that serves every CSP. | 因为我们是唯一服务于每个CSP的公司. |
| 827 | We serve every NCP. | 我们为每一个全国大会党服务。 |
| 828 | We have great enterprise businesses through Dell and HP and Lenovo and others. | 我们通过Dell和HP以及Lenovo等公司有伟大的企业业务。 |
| 829 | Because we serve every single industry and work across the entire life cycle of AI, | 因为我们服务于每一个行业 并在整个AI的生命周期工作, |
| 830 | everything from, of course, data processing, to training, to post-training, to inference, | 当然,从数据处理到培训,到培训后,到推断, |
| 831 | to agentic processing, and all of the tool use that we accelerate. | 用于代理处理, 所有的工具我们加速。 |
| 832 | Our exposure to our go-to-market is the largest of all. | 我们上市的机会最多 |
| 833 | Every contract manufacturer in the world is working on my behalf. | 世界上每个合同制造商都代表我工作 |
| 834 | Every go-to-market is connected to architecture. | 每个上市都与建筑有关. |
| 835 | And so we have one particular state, one particular community, | 所以,我们有一个特定的州, 一个特定的社区, |
| 836 | one particular data center has a hard time. | 一个特定的数据中心有一个困难时期。 |
| 837 | Our demand is so much greater than our supply at the moment. | 我们的需求比现在的供给大得多。 |
| 838 | We'll just adjust our supply chain accordingly. | 我们只需相应调整我们的供应链. |
| 839 | The other thing that we have, and I don't love this answer as much, | 还有一件事,我不爱这个答案, |
| 840 | but as you know, NVIDIA is also the infrastructure for the planet. | 但如你所知 NVIDIA也是地球的基础设施 |
| 841 | And so whether it's in Europe, and we've got early developments now in Africa, | 因此,无论它在欧洲, 我们已经有了早期的发展 现在在非洲, |
| 842 | in Central Asia, in Southeast Asia, in Latin America, in Middle East, and so on and so forth. | 分布于中亚,东南亚,拉丁美洲,中东等地. |
| 843 | And so I don't love that the United States can't keep up. | 所以我不喜欢美国跟不上 |
| 844 | But for our business specifically, we have such a global reach. | 但具体地说,对于我们的事业来说,我们具有如此广泛的全球影响力。 |
| 845 | And there's build-out demand all around the world. | 并且世界各地都有建设需求. |
| 846 | If we don't build it out here, I think we just all have to recognize that every country in the | 如果我们不在这里建造它, 我想我们只要承认 |
| 847 | world has the potential to do that. | 世界有这样做的潜力。 |
| 848 | So I think that our business specifically is probably the most resilient. | 所以,我认为我们的业务 很可能是最有弹性的。 |
| 849 | And the reason for that is all the things that I mentioned. | 原因就是我提到的一切。 |
| 850 | We cover the entire spectrum of AI. | 我们覆盖了AI的整个范围. |
| 851 | We cover the entire spectrum of computing. | 我们涵盖整个计算范围。 |
| 852 | We're in every single cloud, every single region. | 我们在每一个云层,每一个区域。 |
| 853 | And so the resilience of our company is much greater. | 因此,我们公司的复原力要大得多。 |
| 854 | But I do think that land, power, and shell is going to be a challenging part next year. | 但我确实认为,土地、权力和贝壳明年将是一个具有挑战性的部分。 |
| 855 | I guess what I hear from the street is there's this idea that maybe there's 50 gigawatts of | 我猜我从街上听到的是 这个想法也许有50千兆瓦 |
| 856 | announced chips that are underway or talked about, and maybe 35 to 40 gigawatts of land, | 宣布的芯片正在进行或讨论, 以及可能35到40千兆瓦的土地, |
| 857 | power, or shell that might come online. | 电源,或者可能上线的壳体。 |
| 858 | So what compute fits into that land, power, and shell and on what timeline | 因此,什么计算适合 土地,权力,和贝壳 和在什么时间线 |
| 859 | is a question up in the air. | 是一个空中的问题。 |
| 860 | But I guess if we think of this through the entire- | 但我想如果我们从整个... |
| 861 | Clark, just to interrupt you there. | 克拉克 打断你一下 |
| 862 | Of course, as you know, NVIDIA's architecture is the most productive. | 当然,你知道,NVIDIA的建筑是最有成果的. |
| 863 | It generates the most tokens per watt. | 它能产生每瓦最大的符号. |
| 864 | And so if the world's amount of watt, power, energy reduces. | 因此,如果世界的瓦特量,电量,能量减少。 |
| 865 | It shifts their choices to the most productive version. | 它将他们的选择转向最有成效的版本。 |
| 866 | And the reason for that is that they still want to make their revenues. | 其原因是他们仍然想要获得收入。 |
| 867 | And so all of my CEO colleagues, they all have high expectations for revenues next year. | 所以,我所有的CEO同事, 他们都对明年的收入抱有很高的期望。 |
| 868 | And they need to deliver it. | 他们需要送货 |
| 869 | Their shareholders are counting on them to deliver it. | 他们的股东要靠他们来交付。 |
| 870 | All their partners need them to deliver it. | 他们的所有伙伴都需要他们提供。 |
| 871 | And so when the amount of power reduces, the quality of the compute you put into it | 所以当电量减少, 计算的质量 你投入其中 |
| 872 | has to be the most extraordinary. | 必须是最不寻常的。 |
| 873 | We're the most productive computing resource you could put in the world today. | 我们是最有生产力的计算资源 你可以在今天的世界。 |
| 874 | We're the most rentable. | 我们是最可租的。 |
| 875 | We're fungible. | 我们是可替换的。 |
| 876 | And we're durable. | 我们很耐用 |
| 877 | And so the forces around the world, because of the shortage of resources, | 因此,由于资源短缺,世界各地的力量, |
| 878 | is shifting interest towards us. | 我们的兴趣正在转移 |
| 879 | Our market share was already very high. | 我们的市场份额已经很高了。 |
| 880 | But surprisingly to everybody, not only are we a large company and our market share is high, | 但令人惊讶的是,我们不仅是一个大公司,而且我们的市场份额很高, |
| 881 | our market share is actually increasing for these reasons. | 由于这些原因,我们的市场份额正在增加。 |
| 882 | And so you're going to see us go from- | 所以你会看到我们从... |
| 883 | And because we offer so much more per gigawatt than we used to, | 因为我们每千兆瓦的价格比以前高很多 |
| 884 | we used to be a GPU company. | 我们以前是GPU公司 |
| 885 | Then we became the largest and fastest growing networking company. | 之后,我们成为规模最大,发展最快的网络公司. |
| 886 | We're now going to be one of the largest CPU companies, surely the fastest growing one. | 我们现在将成为最大的CPU公司之一, 肯定是增长最快的公司。 |
| 887 | So because we're offering more and more, and we're in computing for storage, | 所以,因为我们提供 越来越多的, 我们正在计算存储, |
| 888 | we're computing for the knowledge infrastructure. | 我们正在计算知识基础设施。 |
| 889 | And now we're in security with Open Shell and Sentry. | 现在我们与开放壳牌和哨兵处于安全状态. |
| 890 | And so the offering that we offer into 1 gigawatt went from, I guess, about 18 billion or so per | 因此,我们提供的1千兆瓦的报价, 我猜,大约180亿左右 |
| 891 | gigawatt to 20-some-odd billion per gigawatt to now 40-plus per gigawatt. | 千兆瓦至每千兆瓦20亿多克,现在每千兆瓦40倍以上。 |
| 892 | And my sense is that that's going to keep on going up. | 我的感觉是,那会继续上升。 |
| 893 | And so not only are we growing share because people are shifting towards our computing platform, | 不仅因为人们正在转向我们的计算平台 而我们的份额越来越大 |
| 894 | we're also growing share because we offer more to the world's CapEx. | 我们还在增加份额 因为我们向世界CapEx提供更多 |
| 895 | And to the world, it's a dollar share of the market, yes. | 对世界来说,市场份额是一美元,是的。 |
| 896 | Yeah. | 对 |
| 897 | If you look at, just go back in just six years ago, | 如果你看看,仅仅在六年前, |
| 898 | our percentage of the world's CapEx was in single digits. | 我们占世界CapEx的百分比是一位数字。 |
| 899 | Then I think right now with the world's CSPs, we're probably running about 40%. | 然后,我想现在 与世界的CSP, 我们可能运行约40%。 |
| 900 | But we're also expanding the market beyond the CSPs. | 但我们也在扩大市场 超越CSP。 |
| 901 | 而CSP之外,无论是Prem还是Neclouds, 我们一半的业务, | |
| 902 | our percentage of CapEx is probably, what, 80%? | 我们的CapEx比例大概是,什么,80%? |
| 903 | And so between half of our business being 40%, half of our business being 80%, | 我们一半的生意是40% 一半的生意是80% |
| 904 | and both of us growing doubling every year. | 并且我们每年都要加倍 |
| 905 | And so you could see that our percentage of the world's CapEx is, in fact, even growing. | 所以,你可以看到,我们的比例 世界的CapEx,事实上,甚至正在增长。 |
| 906 | One of those buyers of Nvidia is somebody we talked about two years ago, Elon, | Nvidia的一个买家 是两年前我们谈过的人 伊隆 |
| 907 | 当他开始建造Corossus。 | |
| 908 | And we were blown away a couple of years ago | 我们几年前就被炸飞了 |
| 909 | to see him stand up a couple of hundred thousand hoppers in Memphis. | 看着他在孟菲斯 站了几十万头黑猩猩 |
| 910 | And I remember you saying nobody can do it like him. | 我记得你说过 没人能像他那样干 |
| 911 | Now he's raising another 40 billion, | 现在他又筹了四百亿 |
| 912 | going to buy maybe a million plus Vera Rubens. | 可能要买一百万加薇拉鲁本斯 |
| 913 | He's moved from the case to the end. | 他从案移到终. |
| 914 | He says he wants to have 10 gigawatts by the end of 2027. | 他说他希望到2027年底能有10千兆瓦. |
| 915 | That's hundreds of billions spent on Vera Ruben. | 即以百千亿那由他佛剎微尘数世界. |
| 916 | It's an unprecedented ramp at scale. | 规模空前的斜坡. |
| 917 | You were with him again yesterday when he was also presented with a medal. | 昨天你又和他在一起,当时他还得到了一枚勋章。 |
| 918 | Is Elon that much better at building compute today as he was then? | 伊隆今天能像当时那样计算得更好吗? |
| 919 | Is it now obvious to you? | 你现在看清楚了吗? |
| 920 | Is it a foregone conclusion that SpaceX will be the largest AI hyperscaler in the world? | SpaceX会不会是世界上最大的AI超标器? |
| 921 | Well, we used to, in the two parts of our chart, Elon used to be | 嗯,我们曾经,在我们图表的两个部分, 伊隆曾经是 |
| 922 | under Enterprise. | 在企业号下。 |
| 923 | That's the way we reported in our numbers. | 这就是我们报告的数字。 |
| 924 | Now he's part of our CSP. | 现在他加入了我们的CSP |
| 925 | Of course. | 当然 |
| 926 | They're massive. | 他们是巨大的。 |
| 927 | I'm delighted that I love working with Elon. | 我很高兴我喜欢和伊隆合作 |
| 928 | He's using our technology for his own AI | 他用我们的技术做自己的人工智能 |
| 929 | in training models and such, deploying models, of course. | 当然,在培训模式和此类,部署模式。 |
| 930 | Grokbot is | 格罗克博特是 |
| 931 | super cool. | 超级酷。 |
| 932 | 对Grokbot的需求 对Grokbot的消费 正在猛增 | |
| 933 | Then of course, he rents it. | 当然是他租的 |
| 934 | The combination of these three things, | 这三件事的结合, |
| 935 | he's fairly unique in doing that. | 他是相当独特的 做到这一点。 |
| 936 | I guess Google does that. | 我猜谷歌会这么做的 |
| 937 | That would be yet another example. | 这将是另一个例子。 |
| 938 | I think he also has the advantage that he loves the intersection between computing | 我觉得他也有他喜欢计算之间的交汇点的优势 |
| 939 | and the physical build-outs. | 和身体积聚。 |
| 940 | He loves that whole area and nobody knows more about energy than he | 他爱那个地区,没有人比他更了解能量 |
| 941 | does. | 当然 |
| 942 | One particular new market that we're going to go explore together is putting data centers out | 我们一起探索的一个新市场 就是建立数据中心 |
| 943 | in space. | 在空间。 |
| 944 | Our two engineering teams are working super closely together to create a new type of | 我们两个工程团队正密切合作 打造一种新型的 |
| 945 | server, a new type of architecture that might be super good in space and on Earth. | 服务器,一种在太空和地球上可能超好的新型建筑. |
| 946 | We saw that Starship launched last week, had its first orbital rendezvous and was launching | 我们看到星舰上周发射 第一次轨道会合 正在发射 |
| 947 | those V3 Sats that literally captured the imagination and broke Twitter for a moment. | 那些V3赛特 字面上吸引了想象力 打破了微博片刻。 |
| 948 | Just people mesmerized by watching these launch into space. | 看着这些飞船飞到太空 让人迷惑 |
| 949 | I can't wait to be a payload. | 我迫不及待想成为载荷 |
| 950 | I mean, not personally. | 我的意思是,不是个人。 |
| 951 | Do you have increasing confidence that we will have a material amount of compute in space | 你是否越来越相信 我们将会在太空中大量计算 |
| 952 | via Starship by let's call it 2028, 2029? | 通过星舰让我们称之为2028,2029? |
| 953 | Elon's put some big numbers out there, | 伊隆把一些大数字 在那里, |
| 954 | but do you think that we can get to material inference in space in that time frame? | 但是,你认为我们可以 得到物质推论 在这个时间范围内的空间? |
| 955 | 只有NVIDIA计算. | |
| 956 | Let me tell you why. | 让我告诉你为什么。 |
| 957 | It's not just about the technology. | 不仅仅是技术 |
| 958 | NVIDIA's performance per unit energy is the world's best. | NVIDIA的单位能量性能是世界上最好的. |
| 959 | We call it performance watt. | 我们称之为性能瓦特。 |
| 960 | But the most important thing is this. | 但最重要的是这个。 |
| 961 | The productivity of NVIDIA's compute is so good, | NVIDIA的计算效率好极了 |
| 962 | meaning the return on invested capital is so good that you could put it out in space | 意思是投资资本的回报是那么高 你可以把它放进太空 |
| 963 | and the launch cost is a small percentage of it. | 而发射费用只占其一小部分。 |
| 964 | The launch cost is not inconsequential. | 发射费用并非不必要。 |
| 965 | You're not going to put low cost computing out there. | 你不会把 低成本的计算在那里。 |
| 966 | You're going to put high value computing | 你会把高价值的计算 |
| 967 | out there. | 那边 |
| 968 | We're the highest value computing system you could imagine. | 我们是你能想象的最高值计算系统 |
| 969 | The highest computing | 最高计算 |
| 970 | system in the history of computing. | 计算史上的系统。 |
| 971 | And so we're the perfect fit, really, to launch into | 所以,我们是最合适的,真的,发射进入 |
| 972 | space right now. | 空间现在。 |
| 973 | Irrespective of the launch cost, it'll still be excellent ROI. | 无论发射费用如何,它仍将是出色的ROI. |
| 974 | So do we get 500 megawatts in space by 29? | 那么到29点我们就能得到500兆瓦了? |
| 975 | Well, I'll let him tell you, but I think- | 我会让他告诉你的 但我想... |
| 976 | Is it possible? | 有可能吗? |
| 977 | Elon's launching | 伊隆发射 |
| 978 | 100 launches a year right now. | 现在每年发射100次 |
| 979 | I don't know what exactly it is. | 我不知道这是什么 |
| 980 | It's going to go to multiple a day. | 一天要进行多次 |
| 981 | Yeah, right. | 没错 |
| 982 | And so you could put Vera | 这样你就可以把薇拉 |
| 983 | Rubin's in there and it'll justify it. | 鲁宾在那里,它会证明这一点。 |
| 984 | The demand is there. | 需求就在那里。 |
| 985 | His return on investment is so good on | 他的投资回报是那么好 |
| 986 | it. | 这个 |
| 987 | And so the productivity is excellent. | 因此,生产力是极好的。 |
| 988 | It's fungible, meaning you put a satellite out in space. | 它是可替换的, 这意味着你把卫星 出太空。 |
| 989 | If you don't have a customer to go rent it, it's at idle and you certainly don't want that. | 如果你没有客户去租, 它闲着,你当然不希望。 |
| 990 | And so | 这样 |
| 991 | NVIDIA compute is going to be fully fungible and there'll be lots and lots of renters. | NVIDIA计算会是完全可替换的 并且会有很多的租户。 |
| 992 | We work with every single close model company in the world. | 我们与世界上每一个亲密的模型公司合作。 |
| 993 | We run every single open model | 我们运行每一个公开模式 |
| 994 | in the world. | 在这个世界。 |
| 995 | We have thousands of startups working with us, so enterprise is around and every single | 我们有成千上万的初创企业 与我们合作, 所以企业是周围和每一个 |
| 996 | single industry, and so the computing is completely | 所以计算是完全的 |
| 997 | fungible and is durable. | 可替换且持久。 |
| 998 | You'll sit up there, and the reason for that | 你会坐在上面, 原因 |
| 999 | is because we keep coup to upgrading the software. | 因为我们不断变换软件 |
| 1000 | We're the only hardware company in the world | 我们是世界上唯一的五金公司 |
| 1001 | that you pay us eight years ago, and we're still | 你付我们八年前, 我们仍然是 |
| 1002 | updating your software, new algorithms. | 更新您的软件,新的算法。 |
| 1003 | And so we'll update the software for as long | 所以我们会更新软件 长久 |
| 1004 | as the satellite's up there, and so it's | 卫星就在上面 所以 |
| 1005 | going to be really durable. | 将是非常持久的。 |
| 1006 | And so I think the combination of these three traits, | 所以我认为这三个特征的结合, |
| 1007 | along with the technology, makes it the perfect payload | 与技术,使它成为完美的有效载荷 |
| 1008 | for SpaceX. | 空间X。 |
| 1009 | It costs trillions of dollars to build all this CapEx. | 建造这些CapEx要花费数万亿美元. |
| 1010 | You guys have leveraged your own balance sheet | 你们借了自己的资产负债表 |
| 1011 | to help folks like the neoclouds build that out. | 帮助像新云的人 建立它。 |
| 1012 | In order for finance people to get it to pencil out, | 为了让金融界的人把它拿出来 |
| 1013 | they need to see payback. | 他们需要看到回报。 |
| 1014 | You talked about some companies seeing one year payback, | 你说过有些公司要赔一年 |
| 1015 | $50 billion in the first year, some spot pricing out | 第一年500亿美元,一些现货定价 |
| 1016 | 也许是SpaceX | |
| 1017 | But if I look at the total incremental offtake, | 但是,如果我看看总的增量, |
| 1018 | the revenue of all the labs this year. | 所有实验室今年的收入。 |
| 1019 | So if I add up the two frontier labs, | 如果我把两个前沿实验室加起来 |
| 1020 | oh yeah, get yourself a snack, and we add up | 哦,是的,给自己一个小吃, 我们加起来 |
| 1021 | some of the inference labs, it's about $200 billion. | 一些推断实验室,大约是2000亿美元。 |
| 1022 | We need to see that double next year to $400 billion. | 我们需要看到明年的两倍到4000亿美元。 |
| 1023 | We need to see it double the year after that to $800 billion. | 我们需要看到这一年翻一番,达到8000亿美元。 |
| 1024 | In order to get in the zip code of the payback | 为了得到回扣的拉链码 |
| 1025 | 以万亿计的CapEx, | |
| 1026 | we had this massive unlock this year with Claude Cohn | 今年我们和克劳德·科恩 一起解锁了这个大锁 |
| 1027 | that led to a 10x increase in anthropics revenues. | 导致人类学收入增加10倍 |
| 1028 | What's your confidence that the labs can continue | 你对实验室可以继续的信心是什么? |
| 1029 | to 2 to 3x their revenues every year | 每年收入2至3倍 |
| 1030 | in order that we get the payback that's | 为了得到回报 |
| 1031 | going to be required to build out | 需要建造 |
| 1032 | 几万亿美元CapEx? | |
| 1033 | I mean, you're investing in these companies. | 我是说 你在投资这些公司 |
| 1034 | You have big investments in these companies, right? | 你对这些公司投资很大,对吧? |
| 1035 | Do you see that level of growth continuing | 你看到这种增长水平在继续 |
| 1036 | in these companies? | 这些公司? |
| 1037 | Oh yeah. | 对 |
| 1038 | People misunderstand what's going on. | 人们误解了发生了什么。 |
| 1039 | So first of all, particularly, I mean, | 所以首先,特别是,我的意思是, |
| 1040 | NVIDIA represents, call it 80% of the revenue share | NVIDIA代表,称之为收入份额的80% |
| 1041 | (原始内容存档于2019-09-21). Introduction of AI basicity in the world today. | |
| 1042 | And so my description of the universe of computing, | 所以我对计算宇宙的描述 |
| 1043 | NVIDIA computing matters here. | NVIDIA计算在这里很重要。 |
| 1044 | They really basically call it three segments, | 他们基本上称之为三段 |
| 1045 | three large segments. | 三个大块 |
| 1046 | It's probably slightly more than that, | 可能不止这些 |
| 1047 | but let me just kind of categorize | 让我来分类一下 |
| 1048 | it in three large segments. | 分为三部分 |
| 1049 | The first segment of our computing | 我们计算的第一个环节 |
| 1050 | is just baseline computing, just meaning | 只是基线计算而已,只是意思 |
| 1051 | used for data processing, molecular dynamics, | 用于数据处理、分子动力学、 |
| 1052 | physics simulation, fusion simulation, | 物理模拟 核聚变模拟 |
| 1053 | fission simulation, you name it, right? | 裂变模拟,你说的对吗? |
| 1054 | Quantum chemistry, quantum physics, and so everything. | 量子化学,量子物理,等等 |
| 1055 | The world of simulation and data processing, | 模拟和数据处理的世界, |
| 1056 | not to mention computer graphics. | 更不用提计算机图形了。 |
| 1057 | And so all of that processing done on our GPUs, | 所以所有的处理 我们的GPU, |
| 1058 | if you just kind of said there was no AI at all, just | 如果你只是说 根本没有AI,只是 |
| 1059 | 古典机器的学习,XGBoost at Visa, | |
| 1060 | my master card for fraud detection, | 我调查欺诈的卡片 |
| 1061 | recommender systems to figure out | 建议系统 |
| 1062 | what news to recommend to you, just all of that stuff | 什么消息给你推荐, 只是所有的东西 |
| 1063 | combined, classical computing, not AI, not deep learning, | 结合,古典的计算,不是AI,不是深刻的学习, |
| 1064 | not agentic systems, I'm going to call it | 不是代理系统,我要称之为 |
| 1065 | a third of our business. | 我们的三分之一生意。 |
| 1066 | And that business is growing, call it 30%, 40% per year. | 这个企业正在增长,叫它每年30%,40%。 |
| 1067 | In the absence of any AI labs, NVIDIA is not zero, right? | 在没有任何AI实验室的情况下,NVIDIA不是零,对不对? |
| 1068 | We're quite a large company. | 我们是一个很大的公司。 |
| 1069 | And so that's number one. | 所以这是第一。 |
| 1070 | The second part is the AI labs, people see two, | 第二部分是人工智能实验室 人们看到两个 |
| 1071 | but we work with every one of them. | 但我们和他们每个人都合作 |
| 1072 | And Metamuse is doing fantastic, right? | 梅塔缪斯做的很棒,对吧? |
| 1073 | And it's vastly NVIDIA. | 并且它是巨大的NVIDIA。 |
| 1074 | And there were zero last year, or a year before that. | 去年是零, 或之前的一年。 |
| 1075 | There were just recommender systems, | 只有推荐系统 |
| 1076 | and now they're killing it with Metamuse. | 现在他们和梅塔缪斯一起杀了它 |
| 1077 | And of course, Grok didn't exist three years ago. | 当然,格罗克3年前就不存在了 |
| 1078 | We didn't run Gemini before, we run Gemini now. | 我们以前没有管理双子座 我们现在管理双子座 |
| 1079 | And starting this year, we started working with Anthropic. | 从今年开始,我们开始和Anthropic合作。 |
| 1080 | And so if you just kind of look at our compute, | 所以如果你看看我们的计算 |
| 1081 | people see a couple of labs, but I see more than that. | 人们看到几个实验室, 但我看到更多的。 |
| 1082 | I see five, six labs, OK? | 我看到5,6个实验室,好吗? |
| 1083 | And so these labs are all doing great. | 所以这些实验室都很好。 |
| 1084 | And if you told me that that represents | 如果你告诉我这代表 |
| 1085 | not quite one third of our business, | 我们没有三分之一的事 |
| 1086 | let's say a quarter of our compute, fantastic. | 我们算出四分之一,太棒了 |
| 1087 | So let's say that's a quarter. | 因此,让我们假设这是四分之一。 |
| 1088 | 当人们看80 NVIDIA, | |
| 1089 | and we now run the infrastructure of the world, | 我们现在管理着世界的基础设施, |
| 1090 | and I just said 25% of it, maybe 30% of it | 我只是说25%,也许30% |
| 1091 | is the Frontier Labs. | 是前沿实验室 |
| 1092 | Call it a third is just processing in general. | 把它称为第三 只是一般的处理。 |
| 1093 | And another third, this is the part that's really great. | 第三个,这是真正伟大的部分。 |
| 1094 | Because we're at the beginning of a new computing platform | 因为我们正处在一个新的计算平台的开端 |
| 1095 | shift, just a vast number of new startups are coming out. | 移动,只是大量的新的初创企业正在出现。 |
| 1096 | So many great companies. | 这么多伟大的公司。 |
| 1097 | And some of them are, of course, doing consumer stuff. | 当然,有些人在做消费者的事情。 |
| 1098 | One of them really doing great, Noah Instinct. | 其中之一真的做得很好,诺亚Instinct。 |
| 1099 | I was just going to mention that. | 我只是想提一下这一点。 |
| 1100 | Noah's doing great. | 诺亚做得很好 |
| 1101 | To all kinds of physical AI, biological AI, material | 各种物理AI,生物AI,材料 |
| 1102 | sciences AI, Jeff Bezos, this new company, Prometheus, | 科学AI,杰夫·贝佐斯,这个新公司,普罗米修斯, |
| 1103 | they're doing fantastic. | 他们做得很好。 |
| 1104 | So the number of companies across the spectrum | 因此,公司的数量 横跨范围 |
| 1105 | of industries that have been started, thousands of them. | 数以千计的工业。 |
| 1106 | And what do they do? | 他们做什么的? |
| 1107 | 他们租NVIDIA电脑。 | |
| 1108 | So they're renters. | 所以他们是租来的 |
| 1109 | And so what is a third of our business | 我们的三分之一生意是什么 |
| 1110 | is kind of cloud computing for the next generation. | 是下一代的云计算 |
| 1111 | It's kind of like AWS for today. | 今天有点像AWS |
| 1112 | And so it's just not AWS for the future. | 所以它只是不是AWS 对于未来。 |
| 1113 | And that represents about a third of our business | 这占我们业务的三分之一 |
| 1114 | altogether. | 完全没有。 |
| 1115 | My sense is that we're in the early innings of the build out | 我的感觉是,我们是在 早期的建设 |
| 1116 | of this new computing architecture. | 这个新的计算结构。 |
| 1117 | And it's a real big mistake to recognize | 认错是大错特错 |
| 1118 | just a couple of companies and think that they're | 只是几家公司 认为他们是 |
| 1119 | funding the infrastructure. | 资助基础设施。 |
| 1120 | And then one observation, open models. | 然后一个观察,开放模型。 |
| 1121 | You know that I've been a great advocate of open models. | 你知道我是公开模特儿的好倡导者 |
| 1122 | And the reason for that is because open models opens up | 原因是开放的模型打开 |
| 1123 | new opportunities. | 新的机会。 |
| 1124 | If you look at all these startup companies, 80% of them | 如果你看这些创业公司,80% |
| 1125 | use open models. | 使用开放式模型。 |
| 1126 | And they're not using it to compete with the frontier labs. | 他们不会用它和前沿实验室竞争 |
| 1127 | They're using it to unlock a new opportunity somewhere else. | 他们利用它打开 一个新的机会在别的地方。 |
| 1128 | It could be in physical AI. | 可能是在身体上的人工智能里 |
| 1129 | It could be manufacturing. | 它可能是制造的。 |
| 1130 | It could be health care. | 这可能是保健。 |
| 1131 | They're doing something completely, completely | 他们正在做一些完全,完全 |
| 1132 | different. | 不一样 |
| 1133 | And they're taking these open models and many of them | 他们正在采取这些公开模型 和许多 |
| 1134 | taking Nemotron. | 带走尼莫特隆 |
| 1135 | And these models are then customized | 然后这些模型被定制 |
| 1136 | into super intelligence. | 进入超级智能。 |
| 1137 | And so if you will, I kind of see | 所以,如果你愿意,我有点看到 |
| 1138 | that the general closed labs will | 一般关闭的实验室会 |
| 1139 | be artificial general intelligence. | 成为人工一般智能。 |
| 1140 | And it gets super good. | 它变得超级好。 |
| 1141 | And then you have all these vertical companies | 然后你有所有这些垂直公司 |
| 1142 | that are focused on a particular domain. | 以特定领域为重点。 |
| 1143 | And they have their proprietary data. | 他们拥有自己的专有数据。 |
| 1144 | They teach it to do something very specific. | 他们教它做一些非常具体的事情。 |
| 1145 | And they achieve super intelligence | 他们实现了超级智能 |
| 1146 | for their domain. | 他们的领地。 |
| 1147 | So you don't see, so again, if we go to your compute math, | 所以,你不明白,所以再次, 如果我们去你的计算数学, |
| 1148 | which I think a lot of people are | 我认为很多人都是 |
| 1149 | using to forecast the future revenues of the labs, | 用来预测实验室的未来收入, |
| 1150 | for example, you said, hey, if you stand up | 比如你说 如果你站起来 |
| 1151 | a gigawatt of this compute, you can get $40, $50 billion | 这个计算值的千兆瓦,你可以得到400,500亿美元 |
| 1152 | out the other end. | 在另一边 |
| 1153 | But if we took the number we talked about earlier, | 但如果我们拿了刚才提到的号码 |
| 1154 | 40 gigawatts of compute coming online next year, | 明年有40千兆瓦的计算机上线 |
| 1155 | and if you took $40 billion out the other end, | 如果你把400亿美元 从另一端, |
| 1156 | that would be $1.6 trillion of revenues coming out | 相当于1.6万亿美元的收入 |
| 1157 | the other end of that factory. | 工厂的另一端。 |
| 1158 | But of course, there's something, | 但当然,有些东西, |
| 1159 | it doesn't work that way. | 那样不行 |
| 1160 | At the end of the day, you've got | 最终,你得到了 |
| 1161 | to look at the demand curve and the rate of diffusion | 以查看需求曲线和扩散速度 |
| 1162 | to determine the revenue that's going to come out | 来决定收入会增加多少 |
| 1163 | the other end of the factory. | 工厂的另一端。 |
| 1164 | And what I hear a lot is that people are nervous. | 我听到的是人们很紧张。 |
| 1165 | They've seen over the course of the year | 一年中他们见过 |
| 1166 | maybe the lab's revenues slowing down a little bit. | 也许实验室的收入 放慢了一点。 |
| 1167 | And they're nervous that at some point, | 他们很紧张,有些时候, |
| 1168 | you're going to have the proverbial bubble popping, | 你会有 谚语气泡弹出, |
| 1169 | where we have more supply coming online than we have demanded | 网上供应比我们要求的要多 |
| 1170 | a moment in time, that AI is more like these step functions | 一瞬间,AI 更像这些步骤功能 |
| 1171 | rather than just this parabolic line straight up. | 不只是这个抛物线直上。 |
| 1172 | I guess the question to you is, are you | 我猜你的问题是,你 |
| 1173 | worried about a moment in time where everybody gets nervous | 担心在某一刻 每个人都会紧张 |
| 1174 | about financing the build out, and then | 有关建设资金,然后 |
| 1175 | we don't get the build out? | 我们没有把建筑弄出来? |
| 1176 | Well, all these different things, | 那么,所有这些不同的东西, |
| 1177 | you've just got to break it down. | 你只需要打破它。 |
| 1178 | Break it down back down to first principles. | 拆解为第一原则. |
| 1179 | And the part that people are trying to extrapolate, | 而人们试图推断的部分, |
| 1180 | the $100 billion of ARR, $125 billion of ARR | 价值1 000亿美元,价值1 250亿美元 |
| 1181 | they're trying to extrapolate, is a small part, | 他们试图推断, 是一个小部分, |
| 1182 | not a small part, it's just a part, | 不是一小部分,这只是一部分, |
| 1183 | of the overall infrastructure build out. | 总体基础设施建设。 |
| 1184 | And so the question is, are people | 所以问题是,人是 |
| 1185 | going to continue doing shopping online? | 继续在网上购物? |
| 1186 | So that third of the build out, let's call that a third, | 因此,三分之一的建筑出来, 让我们称之为第三, |
| 1187 | you said it's a third of your business, take that to the bank. | 你说这是你三分之一的生意, 拿到银行。 |
| 1188 | That's not even a debate. | 这甚至不是一个辩论。 |
| 1189 | So a third of the build out is going to hurt. | 所以三分之一的建筑会受伤。 |
| 1190 | I said a few couple, was it a couple years ago, | 我说几对夫妇, 是几年前, |
| 1191 | that as we exit this decade, infrastructure build out | 当我们走出这十年, 基础设施建设 |
| 1192 | will probably be somewhere between $3 to $4 trillion a year. | 每年大概有3到4万亿美元左右 |
| 1193 | Obviously, it's closer to $4 than it is closer to $3. | 很明显,比起3美元它更接近4美元. |
| 1194 | Per year. | 每年。 |
| 1195 | Per year, yeah. | 每年,是的。 |
| 1196 | By the end of the decade. | 至十年终. |
| 1197 | Yeah, and if you said that $1.5 trillion of that, | 是啊,如果你说1.5万亿美元, |
| 1198 | let's just compute. | 让我们只是计算。 |
| 1199 | Just compute. | 只是计算。 |
| 1200 | Now, why is it going to be so big? | 为什么会这么大? |
| 1201 | Well, it was half a trillion before AI started. | 好吧,这是半兆 之前AI开始。 |
| 1202 | Right. | 对 |
| 1203 | And if I told you that over the course of some six, seven, | 如果我告诉你, 在大约六,七的过程中, |
| 1204 | eight years, it became $1.5 trillion | 八年,变成一万五千亿 |
| 1205 | because the world is just consuming more compute, | 因为世界正在消耗更多的计算, |
| 1206 | I mean, that's not illogical. | 我的意思是,这不是不合逻辑的。 |
| 1207 | On top of this, remember, Moore's Law is over. | 除此之外,记住,摩尔定律已经结束了. |
| 1208 | There was a time, if you want to get twice as many transistors, | 曾经有过, 如果你想得到两倍的晶体管, |
| 1209 | you just wait one year. | 你只要等上一年时间 |
| 1210 | And if you want twice as many transistors, | 如果你想要两倍的晶体管, |
| 1211 | get another, wait one year. | 再拿一个,等一年 |
| 1212 | Same price. | 同样的价格。 |
| 1213 | But if you want twice as many transistors now, | 但如果你现在想要两倍的晶体管 |
| 1214 | doesn't matter how long you wait, | 无所谓你等了多久 |
| 1215 | you have to pay twice as much, OK? | 你得付两倍的钱,好吗? |
| 1216 | Because Moore's Law is over. | 因为摩尔定律结束了. |
| 1217 | And so whatever computing capacity that the world wants, | 所以不管世界需要什么计算能力 |
| 1218 | every single year, if you want twice as much, | 每年,如果你想要两倍, |
| 1219 | you're going to pay twice as much. | 你要付两倍的钱 |
| 1220 | And so on top of that half a trillion dollars that | 因此,在半兆美元之上, |
| 1221 | was kind of growing just naturally | 那种生长 只是自然 |
| 1222 | because the world just used more and more digital, | 因为世界只是使用 越来越多的数字, |
| 1223 | you're going to have to effectively give it an X factor | 你得给它一个X因素 |
| 1224 | just to account for the fact that Moore's Law is over. | 就是为了说明摩尔定律结束了 |
| 1225 | And so without AI, the semiconductor industry | 所以没有AI,半导体工业 |
| 1226 | was on its path to become the largest industry in the world. | 正在走向成为世界上最大的工业。 |
| 1227 | And I've mentioned that long ago, just people forgot. | 我早就提到过,只是人们忘记了。 |
| 1228 | This is one of those things that I said. | 这是我说过的话之一。 |
| 1229 | I just said that two years ago, yeah. | 两年前我刚说过,是的。 |
| 1230 | So that's just one part. | 所以这只是一部分。 |
| 1231 | And so the only thing that we, and then don't forget, | 所以只有一件事,我们,然后不要忘记, |
| 1232 | because we're in a new computing platform shift, | 因为我们在新的计算平台转变, |
| 1233 | there are going to be a whole lot of startups. | 有很多的创业者 |
| 1234 | Innovation is part of the economy. | 创新是经济的一部分。 |
| 1235 | New companies come. | 新公司来了 |
| 1236 | And we said it's accelerating. | 我们说它正在加速。 |
| 1237 | So if we spent, if you go back to Moore's Law, | 所以如果我们花了, 如果你回到摩尔定律, |
| 1238 | general purpose compute, $5 billion per gigawatt. | 一般用途计算,每千兆瓦50亿美元。 |
| 1239 | 今天一个Vera Rubin是50美元。 | |
| 1240 | So the question is, is Vera Rubin | 所以问题是 薇拉·鲁宾 |
| 1241 | deliver 10x more value in the world? | 给世界带来10x的更多价值? |
| 1242 | The punch line is, I don't think it's necessary to extrapolate | 句子是,我觉得没必要推断 |
| 1243 | any of the labs into multiple trillions of dollars. | 任何实验室都变成数万亿美元。 |
| 1244 | Unnecessary to do that. | 没必要这么做 |
| 1245 | I think that's a, I mean, it's an important one. | 我认为这是一个,我的意思是, 这是一个重要的。 |
| 1246 | Because I think people, we just saw yesterday, | 因为我认为人们, 我们刚刚看到昨天, |
| 1247 | even in the market, some chatter about, | 即使是在市场上, 一些闲聊, |
| 1248 | was OpenAI's revenue $70 billion or $50 billion? | OpenAI的收入是700亿美元还是500亿美元? |
| 1249 | I think you and I- | 我觉得你和我... |
| 1250 | They're gonna be insanely successful companies. | 他们会是疯狂成功的公司。 |
| 1251 | Correct. | 没错 |
| 1252 | But you saw the pullback | 但你看见了撤退 |
| 1253 | in a lot of the infrastructure companies, right? | 在很多基础设施公司,对不对? |
| 1254 | 担心OpenAI没有钱 | |
| 1255 | won't have the money in order to pay for the infrastructure. | 没有钱来支付基础设施。 |
| 1256 | I think the point that you're hammering home | 我认为你正在敲敲回家 |
| 1257 | is that when we look at the 40 gigawatts | 是,当我们看 40千兆瓦 |
| 1258 | of infrastructure built out next year, | 明年建成的基础设施, |
| 1259 | at least a third of it is just going to, right, | 至少有三分之一是要去,对吗, |
| 1260 | general purpose compute replacement, | 通用计算替换, |
| 1261 | the basic stuff out of the hyperscalers. | 基本的东西 出超尺度。 |
| 1262 | That a third of it's going to kind of | 三分之一会有点 |
| 1263 | NeoCloud, Sovereign, Enterprises, | 新云,主权,企业, |
| 1264 | everybody else who's using this, all the startups. | 其他人谁使用这个, 所有启动。 |
| 1265 | And a third of it's going to not two labs, | 三分之一不会有两个实验室 |
| 1266 | but to many labs, right, including Grok | 但很多实验室,包括格罗克 |
| 1267 | and Thinking Machines and all the other- | 思考机器和所有其他东西... |
| 1268 | We're putting way too much pressure on two labs. | 我们给两个实验室压力太大了 |
| 1269 | Yeah. | 对 |
| 1270 | I'm delighted to partner with them and support them. | 我很高兴和他们合作 支持他们 |
| 1271 | They're gonna be two of the most consequential | 他们将会是两个 最大的后果 |
| 1272 | technology companies ever built in history. | 科技公司在历史上曾兴建. |
| 1273 | They're gonna be magnificently successful. | 他们会非常成功 |
| 1274 | But they don't have to be anywhere near a trillion dollars | 但他们不必接近一万亿美元 |
| 1275 | in revenues to be successful. | 收入是成功的。 |
| 1276 | Right, right. | 对,对。 |
| 1277 | And I think we're putting way too much pressure on that. | 我觉得我们给这个压力太大了 |
| 1278 | And then the rest of the industry, | 然后其他的行业, |
| 1279 | we still do have to build out a few trillion dollars | 我们还是得赚几万亿 |
| 1280 | of computing infrastructure. | 计算基础设施。 |
| 1281 | Just to keep going, let's shift a little bit. | 继续走,我们换位 |
| 1282 | How good is this, you guys? | 这有多好,你们? |
| 1283 | By the way, what are these called? | 顺便说一下,这些叫什么? |
| 1284 | These are dried plums. | 这些是干梅子 |
| 1285 | Dried plums, the old dried plums. | 干梅,旧干梅. |
| 1286 | Yeah, this one particular store in Taiwan | 对,台湾的这家店 |
| 1287 | is one of my favorites. | 是我的最爱之一。 |
| 1288 | Kind of a licorice-y. | 有点甘草 |
| 1289 | There is licorice, yeah, licorice. | 有甘草,对,甘草。 |
| 1290 | A little licorice-y. | 一点甘草 |
| 1291 | Salt, sugar, cinnamon. | 盐,糖,肉桂. |
| 1292 | Yeah, that's a little twang to it. | 是啊,这是有点淡。 |
| 1293 | Yeah, yeah, you got it, you know? | 是啊,是啊,你得到了它,你知道吗? |
| 1294 | Yeah. | 对 |
| 1295 | Okay, you said as to next year, | 好吧,你说明年, |
| 1296 | demand for Nvidia's chips, well over 100%. | 对Nvidia的芯片的需求 超过100% |
| 1297 | You guided to 70% revenue growth. | 你引导了70%的收入增长。 |
| 1298 | You gave a year forward look | 你往前看一年了,你给我看一年 |
| 1299 | because you said supply's constrained. | 因为你说供应有限 |
| 1300 | That's all I had at the time. | 我当时只有这些 |
| 1301 | That's right. | 没错 |
| 1302 | I got 18 months to work. | 我还有18个月要工作 |
| 1303 | Right, you said you were very confident. | 你说你很自信 |
| 1304 | I'll come to work every day. | 我每天都会来上班 |
| 1305 | You said very confident and plenty of time | 你说很自信,还有时间 |
| 1306 | to work with the supply chain to do even better. | 与供应链合作,以做得更好。 |
| 1307 | You expected to do even better | 你本来想做得更好 |
| 1308 | and then not disappoint customers. | 而不是让顾客失望 |
| 1309 | So are things going even better? | 事情还顺利吗? |
| 1310 | And has the order book continued to grow? | 订单书是否继续增长? |
| 1311 | Because the other part of that is, right, | 因为另一部分是,没错, |
| 1312 | you took a snapshot at a moment in time | 你一瞬间拍了个快照 |
| 1313 | and of course people are continuing to place orders. | 当然人们在继续下命令 |
| 1314 | So do you think there's enough slack in the supply chain | 所以,你认为供应链有足够的松懈 |
| 1315 | you can squeeze more out? | 你可以挤出更多的? |
| 1316 | And is the order book continuing to grow | 命令书还在继续增长 |
| 1317 | as more of these capabilities are unlocked? | 更多这些能力解锁? |
| 1318 | For the moment, I'll tell you why I told everybody. | 现在,我告诉你为什么我告诉大家。 |
| 1319 | The challenge that we have is we have giant order books. | 我们面临的挑战是,我们有巨型订单。 |
| 1320 | 每个CSP,每一个NCP, | |
| 1321 | every single OEM had already placed orders on us. | 每一个OEM已经给我们下了命令 |
| 1322 | Startups, big companies, small companies, | 创业、大公司、小公司 |
| 1323 | they already placed orders on us. | 他们已经给我们下了命令 |
| 1324 | And the reason for that is because we work with them on it, | 原因是我们和他们合作 |
| 1325 | depending on the size of the customer, | 根据顾客的大小, |
| 1326 | at least a year for the largest ones, two years forecast. | 最大国家至少一年,预测两年。 |
| 1327 | And so we already have a very large order book. | 我们已经有了一本大书 |
| 1328 | And because I was in a process of allocating | 因为我正在分配 |
| 1329 | and what I was allocating to all of our customers | 以及我分配给所有顾客的 |
| 1330 | is a lot less than they wanted. | 比他们想要的要少得多 |
| 1331 | If I didn't publicly say what we had, | 如果我不公开说出我们所拥有的一切 |
| 1332 | everybody would be second guessing, | 每个人都会是第二个猜测, |
| 1333 | are they short-changed, are we being, you know? | 他们是短小的改变, 我们是在,你知道吗? |
| 1334 | And so it was my way of essentially handling | 所以这是我处理的方式 |
| 1335 | every single conversation where I'm disappointing people. | 每一次我让人失望的对话 |
| 1336 | And let's say that- | 这么说吧... |
| 1337 | So pretty much every customer would buy more from you | 几乎每个顾客都会从你那里买更多 |
| 1338 | if you could deliver more. | 如果你能提供更多。 |
| 1339 | By far. | 远曰. |
| 1340 | And then since then, of course, | 从那时起,当然 |
| 1341 | we have new orders almost weekly and daily. | 我们有新的订单 几乎每周和每天。 |
| 1342 | And so in addition, however, we're getting more supply | 此外,我们得到了更多的供给 |
| 1343 | and some of the supply that is limiting us | 和一些限制我们的供应 |
| 1344 | is land power shell downstream. | 下游是陆地动力壳 |
| 1345 | And so a couple of weeks ago, | 所以几个星期前 |
| 1346 | we announced two gigawatts in Australia. | 我们在澳大利亚宣布了两千兆瓦 |
| 1347 | And so that's a very big deal. | 所以,这是一个非常大的问题。 |
| 1348 | Two gigawatts to us is like $80 billion. | 两千兆瓦就等于800亿美元 |
| 1349 | Huge. | 巨大的。 |
| 1350 | And then I'm hoping when I go to Europe with UTC, | 我希望当我去欧洲与UTC, |
| 1351 | we'll be able to announce some new partnerships there | 我们可以在那里宣布一些新的伙伴关系 |
| 1352 | and unlock more land power shell. | 并打开更多的土地动力壳。 |
| 1353 | And so this land power shell is clearly one of the obstacles | 这颗陆地动力壳显然是障碍之一 |
| 1354 | and this is an area where NVIDIA has a huge advantage. | 这是NVIDIA拥有巨大优势的地区 |
| 1355 | Because based on what we build, | 因为我们的建设, |
| 1356 | if there's a NeoCloud and this NeoCloud | 如果有新云和这个新云 |
| 1357 | like Coreweave or Nebius, | 就像科里韦或尼比乌斯 |
| 1358 | and they would like to scale out their business, | 他们想扩大他们的生意, |
| 1359 | NVIDIA有整个堆叠, | |
| 1360 | complete infrastructure. | 完整的基础设施。 |
| 1361 | And so we could support them as their scale out. | 这样我们就可以支持他们,让他们扩大规模。 |
| 1362 | And so we could build in places | 这样我们就可以建造一些地方 |
| 1363 | and we can build with partners and customers | 我们可以与合作伙伴和客户一起建设 |
| 1364 | that nobody else can. | 别人都做不到 |
| 1365 | And so I think this gives us a lot more resilience, | 所以我认为这给了我们更多的弹性, |
| 1366 | but nonetheless, as we continue to increase supply | 然而,随着我们继续增加供应 |
| 1367 | upstream and downstream, | 上游和下游, |
| 1368 | one of the critical areas is substrate. | 其中一个关键领域是基底。 |
| 1369 | Substrates are getting larger and there are more chips. | 底物越来越大,还有更多的芯片。 |
| 1370 | And so it's working against them in two ways. | 因此,这在两个方面对他们不利。 |
| 1371 | Every single chip that's being shipped | 运来的每一个芯片 |
| 1372 | uses more of the material. | 使用更多的材料。 |
| 1373 | And so these organic materials are really hard | 所以这些有机材料真的很难 |
| 1374 | to make it the precision that we need | 让它达到我们需要的精确度 |
| 1375 | because as the packages get larger, | 因为随着包裹的扩大 |
| 1376 | the thermal coefficients and the way we stack things | 热系数和我们堆放东西的方式 |
| 1377 | make it really hard to handle things like warpage | 很难处理像战争页之类的事情 |
| 1378 | and various manufacturing defects. | 和各种制造缺陷。 |
| 1379 | And so at the scale that we do, | 因此,在我们做的规模, |
| 1380 | we have to be very precise | 我们必须非常精确 |
| 1381 | and we need really great partners. | 我们需要真正的伟大的伙伴。 |
| 1382 | And so that's supply constraint. | 这就是供应的制约。 |
| 1383 | Those are just some examples. | 这些只是一些例子。 |
| 1384 | Silicon photonics is supply constrained. | 硅光子供应受到限制。 |
| 1385 | And so everything is supply constrained. | 所以一切都是有限的供给。 |
| 1386 | And so we're gonna keep working on it. | 所以我们要继续努力 |
| 1387 | I have 18 months to do so. | 我还有18个月可以做 |
| 1388 | Every single week that goes by, we increase. | 每过一周,我们就增加。 |
| 1389 | But nonetheless, we keep increasing more supply | 但尽管如此,我们继续增加供应 |
| 1390 | and the order books keeps going up. | 和命令书不断上升。 |
| 1391 | And so it's gonna be hard to close that gap | 所以很难弥补这个缺口 |
| 1392 | for a few years. | 几年时间 |
| 1393 | Take a look at this chart. | 看这个图表。 |
| 1394 | This is like a deposition. | 这就像一个证词。 |
| 1395 | No, that's right. | 不,没错。 |
| 1396 | Excuse me. | 对不起。 |
| 1397 | Now usually it goes like this. | 现在通常是这样的 |
| 1398 | Mr. Huang, please now look at it. | 黄公子请今看. |
| 1399 | Good deposition. | 良好的证词。 |
| 1400 | Okay, so the gray line, this was the consensus | 好吧,所以灰色线,这是共识 |
| 1401 | at the start of 2025. | 在2025年初。 |
| 1402 | This was the consensus for NVIDIA at the start of 2025. | 这是在2025年初对NVIDIA的共识. |
| 1403 | The blue line was altimeters forecast way back then. | 蓝线是高程预报 早在那时。 |
| 1404 | So we were a lot higher than the consensus. | 所以我们比共识高得多。 |
| 1405 | The green line is now where the new consensus | 绿线现在是新的共识所在 |
| 1406 | is based upon your guidance. | 这是基于你的指引。 |
| 1407 | And so one of the things that blows my mind about this | 所以有一件事让我心烦意乱 |
| 1408 | is the 25 people who get paid to just forecast | 是25个人的薪水 刚刚预测 |
| 1409 | what this company is going to do, | 这家公司会怎么做? |
| 1410 | their consensus of all their ideas, right, | 他们的共识 他们的所有想法,正确的, |
| 1411 | was off by a couple hundred percent, right, | 被关闭了百分之百,对不对, |
| 1412 | just at the beginning of 20. | 20岁开始 |
| 1413 | Oh no, it's time last year. | 哦,不,这是时间去年。 |
| 1414 | At this time last year, not even 25. | 去年这个时候,甚至没有25个。 |
| 1415 | This green line? | 绿线? |
| 1416 | Yes, that was the consensus at this time last year. | 是的,这是去年这个时候达成的共识。 |
| 1417 | Okay, one year ago. | 好吧,一年前。 - 好吧,一年前。 |
| 1418 | We did that last week. | 我们上周就这么做了 |
| 1419 | Okay, so in a year, the consensus forecast | 好吧,一年后,共识预测 |
| 1420 | has shifted up by 100 to 200% for this company. | 这家公司已经上调了100到200% |
| 1421 | Extraordinary, like in my 20 plus years of doing this, | 特别的,就像在我的20多年 这样做, |
| 1422 | you don't often see consensus being off by that much. | 你并不经常看到共识 被这么多。 |
| 1423 | And yet the stock's only up 20%, okay? | 但股票只有20%,好吗? |
| 1424 | You literally doubled what the revenue forecast was | 你将收入预测翻了一番 |
| 1425 | and the free cash flow forecast was for the company. | 而免费现金流预测是公司。 |
| 1426 | You're now trading at 13 to 14 times earnings | 你现在以13到14倍的收入交易 |
| 1427 | depending upon who you're looking at. | 看你在看谁 |
| 1428 | So what do you think the market is saying | 你觉得市场在说什么? |
| 1429 | about Nvidia or about the confidence | 关于恩维迪亚或信心 |
| 1430 | in the expectations of the market, right, | 对市场的期望,正确的, |
| 1431 | to cause that to be the case? | 导致这种情况? |
| 1432 | Because usually if you see it shift up by that much, | 因为通常,如果你看到它 转移了这么多, |
| 1433 | the stock would have shifted up by a lot more. | 股票本会增加很多 |
| 1434 | First of all, for a company to be as large as us | 首先,公司要和我们一样大 |
| 1435 | gaining share tells you something about the resilience | 获取股份 告诉你一些关于韧性 |
| 1436 | and the size and scale of our advantage. | 以及我们优势的规模和规模 |
| 1437 | Is that advantage increasing or decreasing? | 这种优势是增加还是减少? |
| 1438 | Increasing, increasing, increasing, increasing, increasing. | 增增增减. |
| 1439 | Your competitive moat is increasing at scale. | 你竞争的护城河正在扩大 |
| 1440 | Yeah, let me give you an example. | 我举个例子 |
| 1441 | Let's just do a thought experiment right now. | 我们现在做个思考实验 |
| 1442 | If you had to hand somebody a $100 billion check | 如果你要给别人一千亿的支票 |
| 1443 | to place a purchase order right now for a next generation. | 给下一代下订单 |
| 1444 | If you want to get in line for Vero Rubin Ultra, | 如果你想为维罗・鲁宾・乌特拉排队 |
| 1445 | it's close to placing order time. | 时间快到了 |
| 1446 | Wow. | 珃 |
| 1447 | Sight unseen. | 看不见 |
| 1448 | Yeah. | 对 |
| 1449 | Right? | 对吧? |
| 1450 | But they have the confidence. | 但他们有自信。 |
| 1451 | And so let's say there's a, | 因此,让我们假设有一个, |
| 1452 | you got to place $100 billion on somebody. | 你得给某人上一千亿美元 |
| 1453 | One. | 一个 |
| 1454 | How do you know it's going to be any good? | 你怎么知道会有什么好事? |
| 1455 | 但因为Nvidia建造它, | |
| 1456 | 你知道这会比Vero Rubin更好 | |
| 1457 | And Vero Rubin's awesome. | 和维罗鲁宾的真棒。 |
| 1458 | And the reason why Vero Rubin, | 而为什么维罗鲁宾, |
| 1459 | the reason why everybody had confidence in Vero Rubin | 所有人都对维罗鲁宾有信心的原因 |
| 1460 | is because Grace Blackwell was awesome. | 因为格蕾丝・布莱克威尔很棒 |
| 1461 | And the world of computing, | 和计算的世界, |
| 1462 | the workloads that these computers do are so complex. | 这些计算机的工作量如此复杂。 |
| 1463 | You can't just have like measure one benchmark. | 你不能像衡量一个基准。 |
| 1464 | Right, right. | 对,对。 |
| 1465 | You know, it's $50 billion. | 你知道,这是500亿美元。 |
| 1466 | Yes. | 对 |
| 1467 | And the number of technologies that have to come together | 以及必须结合的技术数量 |
| 1468 | to deliver the performance is insane. | 表演是疯了。 |
| 1469 | You know? | 你知道吗? |
| 1470 | And so the question is, | 所以问题是, |
| 1471 | sight unseen, | 看不见, |
| 1472 | who can you place a $50, $100 billion order on? | 你能给谁5千亿一千亿的订单? |
| 1473 | Yeah. | 对 |
| 1474 | We're literally the only one. | 我们是唯一一个 |
| 1475 | Yeah. | 对 |
| 1476 | Now, back when it was place a PO, | 现在,当它是一个PO, |
| 1477 | a billion dollar PO, | 10亿波兰比索, |
| 1478 | that was challenging enough. | 这已经够有挑战性的了。 |
| 1479 | But if you had placed a $100 million PO | 但如果你放了一亿波兰元 |
| 1480 | and the system is sitting right in front of you, | 系统就坐在你面前 |
| 1481 | you could benchmark the chips. | 你可以把芯片作为基准 |
| 1482 | Right. | 对 |
| 1483 | Right? | 对吧? |
| 1484 | How do you benchmark these systems? | 你如何衡量这些系统? |
| 1485 | Just building it is a miracle. | 只是建造它是一个奇迹。 |
| 1486 | You know? | 你知道吗? |
| 1487 | So if somebody wants to do an evaluation, | 所以如果有人想做个评估 |
| 1488 | I'm happy to support it, | 我很乐意支持它, |
| 1489 | but you're gonna need to give me a giant data center | 但你得给我一个巨大的数据中心 |
| 1490 | and we're gonna need a couple of years to retrofit it. | 我们需要几年的时间来改造它 |
| 1491 | And you know? | 你知道吗? |
| 1492 | And so you can't even benchmark these things anymore. | 所以,你甚至不能 基准这些东西了。 |
| 1493 | Interesting. | 有意思,这很有趣 |
| 1494 | And so it's exactly what happened with TSMC. | 这正是TSMC发生的情况。 |
| 1495 | The reason why we give 100% of our business to TSMC | 我们把业务的百分之百都交给了TSMC的原因 |
| 1496 | is because they're the only company in the world | 因为他们是世界上唯一的公司 |
| 1497 | that I can plan five years out. | 我可以计划五年 |
| 1498 | Right. | 对 |
| 1499 | And sight unseen, | 视而不见, |
| 1500 | know that that transistor is gonna be really good | 知道那个晶体管会很棒的 |
| 1501 | because it's better than that one, better than that one. | 因为它比那个好 比那个好 |
| 1502 | Yeah. | 对 |
| 1503 | And that they will have capacity for me | 他们将拥有我的能力 |
| 1504 | and that they know how to work with my business team, | 他们知道如何与我的团队合作, |
| 1505 | that they have the agility to handle all of my ups and downs | 他们有敏捷的处理 我的所有升降 |
| 1506 | and you know, for example, | 你知道,比如说, |
| 1507 | to deal with the fact that I went from gray to blue to green. | 面对我从灰色变成绿色的事实 |
| 1508 | Yeah, exactly. | 没错 |
| 1509 | Okay? | 好吗? |
| 1510 | They have to be big enough to help me go | 他们必须足够大 帮助我走 |
| 1511 | from gray to blue to green. | 从灰色到蓝色到绿色。 |
| 1512 | Exactly. | 没错 |
| 1513 | There's no other company in the world | 世上没有别的公司了 |
| 1514 | that can help me do that. | 那能帮助我做到这一点。 |
| 1515 | That's right. | 没错 |
| 1516 | And to have all these different process technologies | 拥有所有这些不同的工艺技术 |
| 1517 | from silicon photonics to logic to memory to push, pull, | 从硅光子到逻辑 到记忆 推拉拉拉 |
| 1518 | push, pull, packaging. | 推,拉,包装。 |
| 1519 | I mean, the number of things that we do is so dynamic. | 我的意思是,我们做的很多事情都充满活力。 |
| 1520 | And so how do you possibly do that with a company | 所以,你怎么能做到这一点与公司 |
| 1521 | that is one-tenth their size? | 那是他们十分之一的尺寸? |
| 1522 | It's impossible. | 这是不可能的。 |
| 1523 | And so I think the big idea is this, | 所以我认为大主意是, |
| 1524 | because we're building infrastructure for the planet | 因为我们在为地球建造基础设施 |
| 1525 | and Nvidia has become an infrastructure company. | 而Nvidia已经成为一家基础设施公司。 |
| 1526 | And because these infrastructures cost so much money | 因为这些基础设施要花很多钱 |
| 1527 | and they're moving and technology is moving so fast, | 他们正在移动 技术在如此快速地移动 |
| 1528 | we're practically, you know, the combination of the trust, | 我们实际上,你知道, 信任的组合, |
| 1529 | the fact that we run everything. | 事实,我们运行的一切。 |
| 1530 | You know, so it started out with some GPT models, | 你知道,所以它开始 一些GPT模型, |
| 1531 | but look at the flash attention mechanisms | 但看看闪光的注意机制 |
| 1532 | that were invented. | 这是被发明的。 |
| 1533 | The way that we parallelized the models | 我们模拟模型的方式 |
| 1534 | and sharded across MVLink 72 and how all of that changed. | 在MVLink 72的全程都变了 |
| 1535 | MOE变了。 | |
| 1536 | And now people are talking about transformer diffusion | 现在人们在谈论变压器的传播 |
| 1537 | models that are hybrid, SSMs that are transformers hybrids, | 混合型,变压器混合型, |
| 1538 | and all these different technologies are all right. | 所有这些不同的技术都是好的。 |
| 1539 | The evolution of all that, we know that it all runs on CUDA. | 所有这一切的演化,我们知道,它都运行在CUDA上。 |
| 1540 | It is the only, we don't know whether you | 这是唯一的,我们不知道你 |
| 1541 | need to do more pre-training, post-training, | 需要做更多的培训前,培训后, |
| 1542 | or just agentic inference. | 或者只是代理推论 |
| 1543 | Well, it turns out we run it all anyways. | 事实证明我们都是运行的 |
| 1544 | And so you could build a $50 billion data center | 这样你就可以建造一个500亿美元的数据中心 |
| 1545 | and know that the fungibility of the work, | 并且知道作品的可互换性, |
| 1546 | and if something were to happen, you know, | 如果有事发生,你知道, |
| 1547 | for whatever reason unknown, you want it to, if you will, | 不知什么原因,你希望它,如果你愿意, |
| 1548 | sublease half of your capacity to somebody, | 转租一半你的能力给某人, |
| 1549 | Nvidia is completely fungible. | Nvidia是完全可替换的。 |
| 1550 | So we're fungible in workload, we're fungible in models, | 所以我们在工作量上可以互换,在模型上可以互换, |
| 1551 | we're fungible across customers, | 我们是互换的客户, |
| 1552 | and we're deployable everywhere in the world. | 我们被部署在世界各地 |
| 1553 | You start a company in Europe, | 你在欧洲开家公司 |
| 1554 | you start a company in United States, | 你在美国创办一家公司 |
| 1555 | but guess what? | 你猜怎么着? |
| 1556 | You would like your customer, | 你会喜欢你的顾客的 |
| 1557 | you want to serve customers in Europe, | 你想在欧洲为客户服务, |
| 1558 | you want to serve customers in Japan. | 你想在日本为顾客服务 |
| 1559 | Well, it turns out we're already there. | 原来我们已经到了 |
| 1560 | Nvidia's infrastructure is already there. | Nvidia的基础设施已经在那里。 |
| 1561 | OpenAI想去欧洲 | |
| 1562 | we'll take them there in a heartbeat. | 我们马上带他们去 |
| 1563 | If they want to go to India, in a heartbeat. | 如果他们想去印度, 心跳。 |
| 1564 | Isn't that right? | 是吗? |
| 1565 | If Anthropic wants to go to Japan, Korea, in a heartbeat. | 如果Anthropic想在心跳中前往日本,韩国. |
| 1566 | We're already there. | 我们已经到了 |
| 1567 | All of these architectures are already there. | 所有这些建筑都已经存在. |
| 1568 | And so I think if you take Nvidia in whole | 所以我想如果你把恩维迪亚整条 |
| 1569 | and then you come back and you say, | 然后你回来说, |
| 1570 | okay, I want to do the thought experiment, | 好吧,我想做思想实验, |
| 1571 | I want to place $100 billion on somebody, | 我想给某个人上一千亿美元 |
| 1572 | how do you do it? | 怎么样? |
| 1573 | A year ago, we were sitting here, | 一年前 我们坐在这里 |
| 1574 | and the consensus forecast was that you were going to be | 协商一致的预测是 你会是 |
| 1575 | flat or shrink in 2027. | 2027年平或缩. |
| 1576 | Now the consensus forecast, you're going to grow 74%. | 现在共识预测,你将增长74%。 |
| 1577 | So I just want to let people let that sit in | 所以,我只是想让人们 让它坐进来 |
| 1578 | because a lot of people follow these consensus forecasts | 因为很多人都遵循这些共识预测 |
| 1579 | and it was really wrong. | 这真的是错误的。 |
| 1580 | They were forecasting. | 他们在预测 |
| 1581 | Did they go back and listen to your podcast from last time? | 他们回去听上次播客的节目了吗? |
| 1582 | You told them. | 你告诉他们了。 |
| 1583 | I explained everything. | 我解释了一切 |
| 1584 | Listen, ladies and gentlemen, I did explain everything. | 听着,女士们先生们,我解释过了 |
| 1585 | And so I look out at 2029. | 所以我看着2029。 |
| 1586 | There was, wasn't it an ROI question? | 有,是不是一个ROI问题? |
| 1587 | Yes. | 对 |
| 1588 | And then there was this custom versus general purpose. | 后来出现了这种习惯与一般目的. |
| 1589 | The reason why Nvidia's general purpose | 恩维迪亚总目的原因 |
| 1590 | is so I can have all this fungibility. | 这样我才能拥有这些可替换性 |
| 1591 | That's right. | 没错 |
| 1592 | The benefit of fungibility is low risk. | 可替换性的好处是风险低。 |
| 1593 | Right. | 对 |
| 1594 | And low risk really comes in to play | 而低风险真的来玩 |
| 1595 | when you're placing big dollars. | 当你放大钱的时候 |
| 1596 | When you're building out 50, | 当你建造50, |
| 1597 | a hundred billion dollars worth of infrastructure, | 价值1000亿美元的基础设施, |
| 1598 | you need to know that that infrastructure | 你需要知道,基础设施 |
| 1599 | is going to be really good for a long time, | 这将是真正的好 在很长一段时间, |
| 1600 | no matter what. | 不管怎样 |
| 1601 | So specialization is the enemy of fungibility. | 因此专业化是可互换性的敌人. |
| 1602 | Specialization is the enemy of durability. | 专门化是耐久性的敌人. |
| 1603 | And so, so long as I can prove | 所以,只要我能证明 |
| 1604 | that our performance is extraordinary, | 我们的表现是非凡的, |
| 1605 | and we prove that every single day, | 我们证明,每一天, |
| 1606 | the moment we can prove that, | 只要我们能证明这一点 |
| 1607 | then fungibility and durability translates to rentability, | 可互换性和耐久性就变成可出租性, |
| 1608 | which translates to productivity. | 也就是生产力。 |
| 1609 | And financeability. | 还有资金保障 |
| 1610 | And therefore, financeability, | 因此,融资能力, |
| 1611 | which leads to the lowest TCO. | 这导致了最低的TCO。 |
| 1612 | Right, right, exactly. | 没错,没错 |
| 1613 | Well, now you're buying back the largest amount in history. | 好吧,现在你买回 历史上最大的数额。 |
| 1614 | And on my math, you could probably buy upwards | 根据我的计算,你也许可以买到 |
| 1615 | of 10% of the company back within the next two years. | 在接下来的两年里 10%的公司。 |
| 1616 | At Goldman, you called Nvidia | 在高德曼,你打电话给恩维迪亚 |
| 1617 | the first growth value company. | 第一增长价值公司。 |
| 1618 | Sounds a lot like Apple's pitch to Warren Buffett. | 听起来很像苹果给沃伦·巴菲特的投手 |
| 1619 | Okay, is that right? | 好吧,是吗? |
| 1620 | Yeah, or more, yeah. | 是的,或者更多,是的。 |
| 1621 | That they could return the capital. | 他们可以归还首都 |
| 1622 | It worked. | 已经成功了 |
| 1623 | Yeah, for sure. | 是的,当然。 |
| 1624 | It worked. | 已经成功了 |
| 1625 | Went from 300, at the time, $30 to 300. | 从当时的300美元到30美元 |
| 1626 | Yeah. | 对 |
| 1627 | 300 plus. | 三百加时. |
| 1628 | And in a lot of ways, | 在许多方面, |
| 1629 | we're going through a similar transition. | 我们正在经历一个类似的转变。 |
| 1630 | For example, the thing I forgot to mention is, | 例如,我忘了提的是, |
| 1631 | and Brad, I think you were about to ask me about that. | 布拉德 我想你正要问我这件事 |
| 1632 | You know, most of our investment | 我们大部分的投资 |
| 1633 | and financing partnership strategies | 伙伴关系战略 |
| 1634 | is trying to set up two business models. | 试图建立两个商业模式。 |
| 1635 | Two new business models. | 两个新的商业模式。 |
| 1636 | One business model we've been working on for some time | 一个我们研究了一段时间的商业模式 |
| 1637 | and will be accelerated in the near future | 并在不久的将来加速 |
| 1638 | is software licensing. | 是软件许可。 |
| 1639 | You know, the body of software that we have | 我们拥有的软件体 |
| 1640 | in our company is just extraordinary. | 在我们公司里是非凡的。 |
| 1641 | And that data center is so large, | 数据中心这么大 |
| 1642 | imagine all the software that runs in it. | 想象一下里面的所有软件 |
| 1643 | And we provide the entire stack. | 我们提供整个堆栈。 |
| 1644 | And our entire stack is completely industrial strength. | 而我们的整个堆积 完全是工业实力。 |
| 1645 | For the CSPs, they want to run their own stack. | 对于CSP来说,他们想自己运行堆栈. |
| 1646 | And so we license or we give them our software open source, | 因此,我们许可 或我们给他们我们的软件开源, |
| 1647 | and then they take it and they modify it | 然后他们拿走它 修改它 |
| 1648 | and they integrate into theirs, | 他们将归于他们, |
| 1649 | and then they operate their services. | 然后他们经营他们的服务。 |
| 1650 | But for a lot of the world's companies, | 但对于世界上很多公司来说, |
| 1651 | they don't have the IT department, | 他们没有信息技术部门, |
| 1652 | they don't have the software department do it, | 他们没有软件部门做, |
| 1653 | so they need me to maintain it. | 所以他们需要我维持它。 |
| 1654 | And so because AI, in the last several years, | 所以,因为AI, 在过去的几年里, |
| 1655 | largely focused on training, | 基本上侧重于培训, |
| 1656 | but now they're moving into deployment, | 但现在他们开始部署 |
| 1657 | they need to operate these as factories. | 他们需要将这些作为工厂运营。 |
| 1658 | And they need to know that I'm on the hook, | 他们需要知道我在钩上, |
| 1659 | I'm on the SLA to support it and maintain it, | 我加入了解放军 支持并维持它, |
| 1660 | keep it secure, keep it safe for as long as we shall live. | 保持它的安全, 保持它的安全,只要我们活着。 |
| 1661 | And so software license for our enterprise business | 因此,软件许可证 我们的企业业务 |
| 1662 | will be very, very large. | 将是非常,非常大。 |
| 1663 | And so, you know, if our software license is, | 所以,你知道,如果我们的软件许可证是, |
| 1664 | you know, just roughly say for simple math, | 你知道,只是粗略地说 简单的数学, |
| 1665 | a couple of thousand dollars per GPU per year, | 每年每个GPU几千美元, |
| 1666 | and we have millions of GPUs, | 我们有数百万GPU, |
| 1667 | we ship, you know, 20 million GPUs, say, a year. | 我们每年运送两千万GPU |
| 1668 | I'm growing, and so half of it goes into world's enterprise, | 我长大了,所以一半 进入世界企业, |
| 1669 | it adds up, okay? | 说来话长 |
| 1670 | And so I'm fairly certain this is going to be | 所以我很确定这是 |
| 1671 | a very big part of our business software license. | 是我们商业软件许可证的一大部分 |
| 1672 | And then, of course, there are some clouds | 当然,还有一些云 |
| 1673 | 我们全靠NVIDIA了 | |
| 1674 | They rely on us for our offtake, our ecosystem. | 他们依靠我们来承担责任 我们的生态系统 |
| 1675 | They also rely on us for maybe as an investment partner | 他们也依赖我们作为投资伙伴 |
| 1676 | in helping them secure their financing and capital. | 帮助他们获得资金和资本。 |
| 1677 | They've secured a land power shell, | 他们已经锁定了一个土地动力壳, |
| 1678 | they're a regional champion, | 他们是地区冠军, |
| 1679 | they've got all kinds of stuff working in their advantage, | 他们有各种东西 在他们的优势, |
| 1680 | but these three areas they need our support on. | 但是这三个领域需要我们的支持。 |
| 1681 | And so in those kind of relationships, | 所以在那种关系中 |
| 1682 | 我们很可能会搬到RefShare | |
| 1683 | And so we'll make an investment, | 所以我们要投资 |
| 1684 | we'll get the benefit of the sale up front, | 我们会得到好处 销售在前面, |
| 1685 | but we'll also have the benefit of RefShare going forward. | 但我们也会从RefShare的前进中获益 |
| 1686 | And so, you know, we're not a cloud service provider, | 所以,你知道,我们不是一个云服务供应商, |
| 1687 | but we're going to be providing offtake | 但我们会提供任务 |
| 1688 | and providing software support for them | 并提供软件支持 |
| 1689 | for, you know, four, five, 10 years. | 因为,你知道,四,五,十年。 |
| 1690 | And so we ought to get some economic return on that. | 所以我们应该得到一些经济回报。 |
| 1691 | 所以RefShare会支持你 | |
| 1692 | What's like app store sale? | 什么是像应用商店销售? |
| 1693 | Like app store sale. | 比如应用商店的销售. |
| 1694 | I just described two new growth vectors | 我刚刚描述了两个新的生长载体 |
| 1695 | independent of the hardware sale of our company, | 独立于我们公司的硬件销售 |
| 1696 | and it's much more related to the install base | 和安装基础有关 |
| 1697 | of our platform. | 我们的平台。 |
| 1698 | What's the benefit of software licensing | 软件许可有什么好处? |
| 1699 | if you have no install base? | 如果你没有安装基础? |
| 1700 | And how could you have RefShare if there's no offtake? | 如果没有责任,你怎么能有RefShare? |
| 1701 | Okay, and so the install base element, | 好吧,那么安装基础元素, |
| 1702 | the richness of our software stack, | 我们软件库的丰富性, |
| 1703 | the offtake, the richness of our offtake | 我们的财富 |
| 1704 | and the ecosystem that we have around the world, | 以及我们在世界各地的生态系统, |
| 1705 | that's going to fuel that. | 那会激化它。 |
| 1706 | One of the ways, you know, | 其中一个方法,你知道, |
| 1707 | so we're talking about this training at 13 times, | 我们13次讨论这个训练 |
| 1708 | you know, when Tim and Luca went to see Warren Buffett, | 你知道,当蒂姆和卢卡去见沃伦巴菲特时, |
| 1709 | and one of the reasons that Buffett has said | 以及巴菲特说的原因之一 |
| 1710 | he doesn't like technology companies | 他不喜欢科技公司 |
| 1711 | is they have this hit, | 他们有这个命中, |
| 1712 | they produce an enormous amount of free cashflow, | 他们产生大量的 自由现金流, |
| 1713 | and rather than ever giving any of it | 而不是永远给它的任何 |
| 1714 | back to the shareholders, | 回到股东, |
| 1715 | they go and spend it on something else. | 他们去花在别的东西上。 |
| 1716 | And oftentimes the something else isn't as good | 经常有别的东西不是那么好 |
| 1717 | as the thing that caused them | 因为它的产生 |
| 1718 | to generate the cashflow in the first place. | 首先生成现金流。 |
| 1719 | You guys are now spending over 50% of your free cashflow | 你们现在要花50%以上的现金 |
| 1720 | buying back your stock, | 买回你的股票, |
| 1721 | effectively returning it, you know, to shareholders | 有效地还给股东 |
| 1722 | while you have this exponential growth at the same time. | 当你同时有这个指数增长。 |
| 1723 | As we've already discussed- | 我们已经讨论过了... |
| 1724 | We said 50. | 我们说50。 |
| 1725 | Right. | 对 |
| 1726 | I think this last quarter we did 60 some on. | 我想这最后一个季度 我们做了60一些。 |
| 1727 | Correct. | 没错 |
| 1728 | Yeah, we're going to increase, | 是啊,我们要增加, |
| 1729 | you know, we invest in really high quality companies | 你知道,我们投资 真正高质量的公司 |
| 1730 | as you guys know. | 如你们所知。 |
| 1731 | And, you know, the last couple of years, | 而且,你知道, 在过去的几年里, |
| 1732 | NVIDIA从谁那里买的, | |
| 1733 | whoever Jensen's buys from, | 詹森从谁那里买的 |
| 1734 | you got to invest in that company. | 你得在那个公司投资 |
| 1735 | That's right. | 没错 |
| 1736 | And now the new word on the street | 现在街上的新词 |
| 1737 | is whoever Jensen invests in, | 谁是詹森投资, |
| 1738 | you got to invest in that company. | 你得在那个公司投资 |
| 1739 | And it's true. | 亦复如是. |
| 1740 | And the reason for that is because, you know, like you, | 原因是,你知道,像你一样, |
| 1741 | we're in the ecosystem, | 我们在生态系统, |
| 1742 | we're in the flow of information, | 我们在信息流通中 |
| 1743 | we know who's doing- | 我们知道是谁做的... |
| 1744 | We do a lot of those investments together. | 我们一起做很多投资。 |
| 1745 | Yeah, and we know a lot of smart people, | 是的,我们认识很多聪明人 |
| 1746 | and we want to back these people | 我们想支持这些人 |
| 1747 | when they want to do something great. | 当他们想做一些伟大的事情。 |
| 1748 | And so we're kind of on the inside, | 所以,我们有点内在, |
| 1749 | we have a lot of information, | 我们有很多信息 |
| 1750 | and so our investments are, you know, well-informed. | 所以,我们的投资,你知道, 完全知情。 |
| 1751 | Right, right. | 对,对。 |
| 1752 | And not to mention, NVIDIA is also in the ecosystem, | 更别提NVIDIA也在生态系统里 |
| 1753 | and so we can help them succeed. | 这样我们就能帮助他们成功。 |
| 1754 | That's right. | 没错 |
| 1755 | And so there's a lot of different reasons | 所以有很多不同的原因 |
| 1756 | why the investments turn out great. | 投资结果如何? |
| 1757 | But as a percentage of our total cash generation, | 但作为我们总现金产生的百分比, |
| 1758 | it's going to decline very rapidly. | 它会迅速下降。 |
| 1759 | 部分原因是OpenAI和Anthropic | |
| 1760 | isn't going to come around again, | 不会再来了 |
| 1761 | that opportunity's not going to come around again, frankly. | 老实说,那机会不会再出现 |
| 1762 | And then secondarily, our cashflow generation | 其次 我们的现金流量 |
| 1763 | is growing quite fast. | 增长很快。 |
| 1764 | Right. | 对 |
| 1765 | And to that point, it seems to me | 至此,我觉得 |
| 1766 | that if you're buying back that much, | 如果你买回这么多, |
| 1767 | and the durability of the revenues, | 和收入的持久性, |
| 1768 | I think, as you've described here today, | 我想,正如你今天在这里所描述的, |
| 1769 | is longer than people expect. | 时间比人们预期的要长 |
| 1770 | And I think that's the one place | 我想这就是一个地方 |
| 1771 | people are still hanging too much on, | 人们仍然挂太多, |
| 1772 | this is only going to last as long, | 这只会持续多久, |
| 1773 | Ezra asked you this on the pod the other day, | Ezra那天在舱里问你这个 |
| 1774 | how long until the bubble bursts? | 多久后泡泡暴发? |
| 1775 | And I think you said something about two to three years | 我想你说过两三年了 |
| 1776 | you have good visibility that the demand is there | 你有一个很好的可见度 需求在那里 |
| 1777 | and who knows kind of after that. | 之后谁知道呢? |
| 1778 | And I heard some people saying, | 我听到一些人说, |
| 1779 | oh, it may only be two years before this burst. | 可能只有两年时间了 |
| 1780 | But when I look at this, | 但当我看到这个, |
| 1781 | this is fundamentally a re-architecture | 这基本上是重构建筑 |
| 1782 | of all global compute forever, forever. | 永远的计算 |
| 1783 | So unless there's a global depression, right? | 除非全球出现抑郁症,对吗? |
| 1784 | Then every single year, people are going to have to consume | 以后每年都有人要消费 |
| 1785 | 大量的NVIDIA计算。 | |
| 1786 | Yeah, and so the way to think about that | 对,所以想这个的方法 |
| 1787 | is we were running in the, | 我们跑在, |
| 1788 | the CSPs were running about half a trillion, | CSP正在运行约半兆, |
| 1789 | the whole industry, the whole world is running, | 整个行业,整个世界都在运行, |
| 1790 | call it a trillion dollars of IT build out each year, okay? | 把它叫做每年一万亿美元 IT建设出来,好吗? |
| 1791 | And every year more gets built, | 每年都会建造更多 |
| 1792 | and every five years or so, | 每五年左右, |
| 1793 | some gets depreciated or turned off and taken offline | 有些会贬值或关机然后下线 |
| 1794 | and so on and so forth, and some of it lasts longer. | 诸如此类,有些会持续更长。 |
| 1795 | And so that old world was in the retrieval way | 所以那个老世界在回收过程中 |
| 1796 | of doing computation and the new way of doing computation, | 以新的方法计算。 |
| 1797 | you know, I've predicted that going out of the decades, | 你知道,我已经预言过 走出几十年, |
| 1798 | so three, four, let's say four, let's say five, | 因此,三,四,让我们说四,让我们说五, |
| 1799 | and my sense is that we're going to continue | 我的感觉是,我们将继续 |
| 1800 | to grow from there. | 从那里成长。 |
| 1801 | If you told me that the annual capacity building, | 如果你告诉我,年度能力建设, |
| 1802 | infrastructure building, refreshing combination | 基础设施建设,更新组合 |
| 1803 | of new and old, refreshes, runs around $10 trillion, | 新旧的,新鲜的, 运行在10万亿美元左右, |
| 1804 | I'm not even a little bit surprised. | 我一点也不惊讶 |
| 1805 | Right, right. | 对,对。 |
| 1806 | Just again, for people listening, | 再说一遍,人们在听 |
| 1807 | you know, you think $4 trillion of annual spend | 你知道,你认为4万亿美元的年度开支 |
| 1808 | by the end of the decade, | 在十年结束时, |
| 1809 | because every time you consume tokens, | 因为每次你食用代币时 |
| 1810 | every time you use intelligence, | 每次你用情报 |
| 1811 | you have to run the machine, the factory has to run. | 你必须运行机器,工厂必须运行。 |
| 1812 | And I think you've described that, that's about 50%, | 我想你已经描述过了 大约50% |
| 1813 | you know, if you assume that, you know, | 你知道,如果你假设, 你知道, |
| 1814 | $50 trillion of global GDP is, you know, | 全球GDP的50万亿美元是,你知道, |
| 1815 | is related to knowledge work. | 与知识工作有关。 |
| 1816 | If we get a 20% productivity gain over five years on that, | 如果我们在五年内得到20%的生产力增益, |
| 1817 | like you've seen inside Nvidia, | 就像你在恩维迪亚看到的那样 |
| 1818 | you've seen much bigger productivity gains inside Nvidia, | 在恩维迪亚,你看到了更大的生产力增长, |
| 1819 | but if you see that sort of productivity gain, | 但是如果你看到这种生产力的提高 |
| 1820 | then if we get $10 trillion of incremental economic activity, | 如果有10万亿的增量经济活动 |
| 1821 | right, by 2029, 2030, like that seems very reasonable, | 对,到2029年,2030年, 似乎非常合理, |
| 1822 | and you're just saying we'll spend half of that on CapEx, | 你只是说我们会把一半的钱花在CapEx上 |
| 1823 | on the intelligence, you know, to produce it. | 情报,你知道的,生产它。 |
| 1824 | And so maybe, you know, that's a good place for us | 所以,也许,你知道, 这是一个好地方给我们 |
| 1825 | to take a bit of an off-ramp. | 以取出一点外侧。 |
| 1826 | You've been incredibly generous. | 你真是太慷慨了 |
| 1827 | I could go on forever, | 我可以永远继续 |
| 1828 | but I kind of want to wrap where we started, right? | 但我想把开始的地方包起来 对吧? |
| 1829 | When we look back. | 回首时. |
| 1830 | If this was one of those eating, drinking podcasts. | 如果这是其中之一 吃,喝播客。 |
| 1831 | We could keep going, we could keep going. | 我们可以继续走, 我们可以继续走。 |
| 1832 | It wasn't so long ago that the human condition | 不久前,人类的状况 |
| 1833 | was described as nasty, brutish, and short, right? | 被描述为肮脏,粗鲁, 和短,对不对? |
| 1834 | Was that last week or? | 那是上周还是? |
| 1835 | Innovation has really driven all human progress, | 创新真正推动了人类的进步 |
| 1836 | mostly in the last 200 years. | 大部分是在过去200年里。 |
| 1837 | More time with family, longer lives, | 更多的时间与家庭, 更长的生命, |
| 1838 | less time doing drudgery | 做工的时间减少 |
| 1839 | that humans really didn't want to do, safer. | 人类真的不想做,更安全。 |
| 1840 | We're the first generation | 我们是第一代 |
| 1841 | where 63-years-olds consider young. | 在那里63岁的孩子认为年轻。 |
| 1842 | Exactly, exactly. | 没错,没错。 |
| 1843 | Exactly. | 没错 |
| 1844 | People call me young man. | 人们叫我年轻人 |
| 1845 | Safer, better lives. | 更安全,更好的生活。 |
| 1846 | All of this, despite all of these concerns | 所有这一切,尽管所有这些关切 |
| 1847 | that you and I hear out there. | 你和我听到那边。 |
| 1848 | And by the way, those same concerns were expressed | 顺便说一句,也表达了同样的关切 |
| 1849 | at the beginning of the industrial revolution, right? | 在工业革命的开始,对不对? |
| 1850 | What's your level of confidence | 你的信任程度是多少 |
| 1851 | that in five years from now, | 五年后,我们开始 |
| 1852 | we will have made the right decisions | 我们将做出正确的决定 |
| 1853 | and the world will be that much better place, right? | 这个世界将会是更好的地方,对吧? |
| 1854 | That we describe because of the investment | 我们描述的是投资 |
| 1855 | and all that we're doing. | 我们所做的一切 |
| 1856 | Are you more confident as we sit here today | 我们今天坐在这里,你更有信心吗? |
| 1857 | than you were 12 months ago? | 比你12个月前还好? |
| 1858 | I'm 99.99% just in service, | 我99.99%是在服务, |
| 1859 | in honor of a good friend, not using zero and 100. | 以纪念一个好朋友, 不使用0和100。 |
| 1860 | I'm 99.99%, 99% confident | 我99.99%,99%有信心 |
| 1861 | that too many of us care due deeply | 太多人关心我们 |
| 1862 | that the outcome turns out well. | 结果是好的。 |
| 1863 | That we will dedicate our might, our intellect, | 我们将奉献我们的力量,我们的智慧, |
| 1864 | our resources to ensure we get there responsibly | 我们的资源确保我们负责任地到达那里 |
| 1865 | and optimistically. | 乐观地说 |
| 1866 | And so there's just so many people doing the right things. | 所以有这么多人做正确的事情。 |
| 1867 | Now, oftentimes, there are people | 现在,经常有人 |
| 1868 | who pontificate about these things | 他们为这些事而造谣, |
| 1869 | and they're informed by several things | 他们从几件事情得知 |
| 1870 | and they don't realize that there are a lot of people, | 他们不知道有很多人 |
| 1871 | industries, incredible researchers and engineers | 工业、不可思议的研究人员和工程师 |
| 1872 | outside their sphere of information | 在其信息范围以外 |
| 1873 | that are doing amazing things all by themselves, | 做惊人的事情 他们自己, |
| 1874 | independent of them, just because they didn't say, | 独立于他们, 只是因为他们没有说, |
| 1875 | they didn't say it doesn't mean a lot of smart people | 他们没说这不代表很多聪明人 |
| 1876 | are doing smart things. | 正在做聪明的事。 |
| 1877 | And I say this all the time inside the company, | 我经常在公司里说 |
| 1878 | even without my leadership, without strategies, | 即使没有我的领导,没有策略, |
| 1879 | the company is doing smart things all by itself. | 公司正在做一些聪明的事 |
| 1880 | Smart people do smart things. | 聪明人做聪明的事 |
| 1881 | They do more impactful things when you organize a line, | 当你组织一条线时,它们会做更多的影响 |
| 1882 | but amazing people are doing smart things | 但令人惊奇的人 正在做聪明的事情 |
| 1883 | all by themselves. | 独自一人 |
| 1884 | And so when somebody tells me that you're worrying about us, | 所以当有人告诉我你在担心我们 |
| 1885 | it's good that we worry. | 我们担心是件好事 |
| 1886 | And we should express our worries. | 我们应该表达我们的担忧。 |
| 1887 | We should express our concerns, | 我们应该表达我们的关切, |
| 1888 | but we shouldn't predict the end of the world. | 但我们不应该预测世界末日 |
| 1889 | Because they're underestimating the human condition. | 因为他们低估了人类的状况. |
| 1890 | They're underestimating the human capacity. | 他们低估了人的能力. |
| 1891 | And they think for some reason with, | 他们认为,出于某种原因, |
| 1892 | maybe it's just hubris, maybe it's arrogance, | 也许只是自大 也许是傲慢 |
| 1893 | that unless they do it, it won't get done. | 除非他们这样做,否则是办不到的。 |
| 1894 | Unless they do it, it won't get done. | 除非他们这样做,它不会完成。 |
| 1895 | The fact of the matter is this company | 事实是这家公司 |
| 1896 | is doing amazing things, whether I say it or not. | 无论我说不说 都在做令人惊奇的事 |
| 1897 | Right. | 对 |
| 1898 | Okay, and I get all the accolades and the blame, | 好吧,我得到所有的优雅 和责备, |
| 1899 | but in both cases, in both cases, I'm fair, | 但是在这两种情况下 我都很公平 |
| 1900 | especially the accolades. | 特别是优等生 |
| 1901 | And so I, and the reason for that is just these, | 所以我 原因就是这些 |
| 1902 | when you surround yourself with amazing people | 当你用神奇的人包围自己 |
| 1903 | and you create the conditions | 然后创造条件 |
| 1904 | by which everybody wants the same outcome, | 每个人都想得到同样的结果, |
| 1905 | the good outcome, the good version of it. | 好结果,好的版本。 |
| 1906 | And I think most people do. | 我认为大多数人都这样 |
| 1907 | That's the reason why I say we need to, | 这就是为什么我说我们需要, |
| 1908 | we need to hold ourselves accountable and responsible, | 我们需要对自己负责和负责, |
| 1909 | but we should express a future that's optimistic. | 但我们应该表达一个乐观的未来 |
| 1910 | For sure. | 当然 |
| 1911 | And I am 99.99% certain that surrounded | 我99.99%的确定率 |
| 1912 | by all the amazing people in the world, | 以全世界所有令人惊奇的人发誓, |
| 1913 | we're gonna get there together. | 我们要一起去 |
| 1914 | 是的,同意了。 \fn微软雅黑Yeah, agreed. | |
| 1915 | And it's, I think you're a national treasure. | 我认为你是国家宝藏 |
| 1916 | We're lucky to have you helping | 我们很幸运你能帮忙 |
| 1917 | to lead the charge. | 领导控罪。 |
| 1918 | The nation's indebted to you for getting everybody together, | 国家多亏你让大家聚在一起 |
| 1919 | getting us all on the same page, building the framework. | 让我们大家同归于尽,构建框架。 |
| 1920 | There's gonna be more work to do. | 还会有更多的工作要做。 |
| 1921 | But as a father of two, and I know you have kids, | 但作为两个孩子的父亲 我知道你有孩子 |
| 1922 | you're gonna have grandkids, | 你会有孙子, |
| 1923 | and we don't care about anything more. | 而且我们不在乎什么了。 |
| 1924 | That's right. | 没错 |
| 1925 | Than actually building a better world for them, right? | 比真正为他们建立一个更好的世界,对不对? |
| 1926 | We could easily stop. | 我们可以轻易停止。 |
| 1927 | But if we stop today, we lose, right? | 但如果我们今天停下来,我们输了,对不对? |
| 1928 | We lose the opportunity at that better world. | 我们失去了在这个更美好世界的机会。 |
| 1929 | So we'll wrap it there. | 我们把它包起来 |
| 1930 | Thank you, Jensen. | 谢谢你,詹森。 |
| 1931 | Thank you, guys. | 谢谢你们 |
| 1932 | And Brad, thank you for, | 布拉德 谢谢你 |
| 1933 | I've never met a greater patriot. | 我从没见过一个更伟大的爱国者 |
| 1934 | Appreciate that. | 谢谢 |
| 1935 | Somebody who cares so much about the industry, | 某个很关心这个行业的人 |
| 1936 | but even more so the country. | 但国家更是这样 |
| 1937 | And your gift, your gift in your enthusiasm | 你的天赋 你的热情 |
| 1938 | and just tireless energy in coordinating, | 并且只是不懈的精力 在协调, |
| 1939 | pulling together a lot of people | 召集很多人 |
| 1940 | in order to do something great like what you did. | 为了做一件和你一样伟大的事 |
| 1941 | Incredible. | 不可思议 |
| 1942 | I like that. | 我喜欢 |
| 1943 | As a reminder to everybody, | 为了提醒大家 |
| 1944 | just our opinions, not investment advice. | 只是我们的意见 不是投资建议 |
| 1945 | I like that. | 我喜欢 |
该视频共有字幕 1951 条。解锁更多字幕为会员功能，请移动到 价格