有钱也不卖！黄仁勋硬核发声：绝不出售芯片给不负责任的 AI
字幕摘录
| 时间 | 英文 | 中文 |
|---|---|---|
| 0:00 | You're traveling across Europe promoting this idea of sovereign AI. | 你穿越欧洲宣传这个主权AI的想法. |
| 0:03 | Is it still sovereign if it's built on American chips? | 如果它是在美国芯片上建造的,它仍然拥有主权吗? |
| 0:06 | Yes. | 对 |
| 0:08 | Sovereignty basically means that you have to control your own experience. | 主权基本上意味着你必须控制自己的经验。 |
| 0:11 | And in the case of AI, that means control your own data. | 就人工智能而言,这意味着控制你自己的数据。 |
| 0:13 | Your data, your life experience, your company experience, your country's culture, values, | 你的数据,你的生活经验,你的公司经验, 你的国家的文化,价值, |
| 0:21 | the institutional knowledge of your people, the flywheel of creating more data, more experience, | 创造更多的数据 更多的经验 |
| 0:28 | more intelligence. | 更聪明一点 |
| 0:29 | That belongs to you. | 这是属于你的。 |
| 0:31 | And the computer is just an instrument. | 而电脑只是一个仪器。 |
展开字幕全文（102 条）
| 序号 | 英文 | 中文 |
|---|---|---|
| 1 | You're traveling across Europe promoting this idea of sovereign AI. | 你穿越欧洲宣传这个主权AI的想法. |
| 2 | Is it still sovereign if it's built on American chips? | 如果它是在美国芯片上建造的,它仍然拥有主权吗? |
| 3 | Yes. | 对 |
| 4 | Sovereignty basically means that you have to control your own experience. | 主权基本上意味着你必须控制自己的经验。 |
| 5 | And in the case of AI, that means control your own data. | 就人工智能而言,这意味着控制你自己的数据。 |
| 6 | Your data, your life experience, your company experience, your country's culture, values, | 你的数据,你的生活经验,你的公司经验, 你的国家的文化,价值, |
| 7 | the institutional knowledge of your people, the flywheel of creating more data, more experience, | 创造更多的数据 更多的经验 |
| 8 | more intelligence. | 更聪明一点 |
| 9 | That belongs to you. | 这是属于你的。 |
| 10 | And the computer is just an instrument. | 而电脑只是一个仪器。 |
| 11 | And so that's one. | 所以这是一个。 |
| 12 | The second is sovereignty means that nobody could deny you service. | 第二,主权意味着没有人可以拒绝你提供服务。 |
| 13 | You know, like, AI is like infrastructure. | 你知道,像,AI就像基础设施。 |
| 14 | It's like electricity. | 如电. |
| 15 | It's like the Internet. | 如来网. |
| 16 | 如 water水. | |
| 17 | And those are critical resources and critical infrastructure for your society and your companies. | 这些是你们社会和公司的重要资源和基础设施。 |
| 18 | You can't ever have a denial of service. | 你永远不能拒绝服务 |
| 19 | And so you have to control that. | 所以,你必须控制它。 |
| 20 | That's sovereignty. | 这就是主权。 |
| 21 | Sovereignty means that you have the ability to determine your own destiny. | 主权意味着你有能力决定自己的命运. |
| 22 | And you might need teamwork, you might need collaboration, you might need partnership | 你可能需要团队合作,你可能需要合作,你可能需要合作 |
| 23 | and support from other people, but at some, you know, you have agency. | 和别人的支持, 但在某些,你知道,你有代理。 |
| 24 | More than 100 data center projects have been blocked or delayed in the U.S. just so far | 超过100个数据中心项目 在美国被封锁或推迟 仅仅到目前为止 |
| 25 | this year. | 今年, |
| 26 | There's pushback in Europe, too. | 欧洲也有后推. |
| 27 | You have AI POs getting delayed, getting scrapped. | 你有AIPO被拖延, 被废掉。 |
| 28 | Are we seeing cracks in the AI boom? | 我们在AI爆炸中看到裂缝吗? |
| 29 | And could you be left holding the bag? | 你能把包放一边吗? |
| 30 | The capabilities of AI, now that we know what it can do, that's never going to go back. | AI的能力,既然我们知道它能做什么,那就永远回不去了. |
| 31 | So I think, I think a lot, many ways it's fundamental and it's irrespective of near-term, | 所以我想,我认为很多, 许多方式,它是根本性的, 它不管短期, |
| 32 | near-term challenges. | 近期挑战。 |
| 33 | Now, the near-term challenges are somehow also self-imposed. | 现在,短期挑战在某种程度上也是自我强加的。 |
| 34 | I think we're, you know, as an industry, we're doing a horrible job associating this technology, | 我认为我们,你知道,作为一个行业, 我们正在做一个可怕的工作 连接这种技术, |
| 35 | which is built by people. | 由人们建造。 |
| 36 | It's engineered, obviously. | 很明显是设计出来的 |
| 37 | We obviously know enough about how it works to make it better every day. | 我们显然已经足够了解如何每天改善状况。 |
| 38 | And because we're building it, we have every possibility to advance it safely and responsibly. | 因为我们正在建造它, 我们完全有可能安全和负责任地推进它。 |
| 39 | And so I think, in a lot of ways, the narrative of the industry is really out of control and | 所以我想,从很多方面来说, 行业的叙述真的失控了, |
| 40 | ought to be stopped. | 应该停止。 |
| 41 | But that's just an industry growing up, you know, and you probably remember the good old | 但这只是一个行业 成长,你知道, 你可能记得好的老 |
| 42 | days of the Internet, and the industry grew up. | 互联网时代, 和产业成长。 |
| 43 | And I think that our industry is, many of the young companies are going through their | 我认为,我们的工业, 许多年轻的公司正在经历他们的 |
| 44 | growing pains and they're growing, of course, incredibly fast, which I'm very proud of. | 越来越痛苦,他们正在增长, 当然,惊人的快速, 我非常自豪。 |
| 45 | They're just incredible engineers and really bright computer scientists. | 他们只是不可思议的工程师和非常聪明的计算机科学家。 |
| 46 | There's an accountability, responsibility that goes along with building companies that | 有问责制,责任 与建设公司 |
| 47 | I think they're just growing into. | 我觉得他们正在成长 |
| 48 | 你是Anthropic和OpenAI的主要投资人. | |
| 49 | These are two of your major customers as well. | 这些也是你的两大客户 |
| 50 | What happens when the Frontier Labs go public? | 边疆实验室对外开放后会怎样?. |
| 51 | How much is at stake and what do they need to prove? | 事关多少,他们需要证明什么? |
| 52 | What I would say is that I'm delighted to see their success. | 我想说的是 我很高兴看到他们成功 |
| 53 | First of all, these are the two fastest growing technology companies in history. | 首先,这是历史上发展最快的两家技术公司。 |
| 54 | Their revenues are real. | 他们的收入是真实的。 |
| 55 | And their profits generated from the revenues are real. | 他们从收入中获得的利润是真实的。 |
| 56 | This last year, this last six months, AI went through a transition. | 去年,过去六个月,大赦国际经历了一个过渡。 |
| 57 | It used to be unprofitable, unuseful, very curious, very great potential, and now we're | 它曾经是无利可图,无用, 非常好奇, 非常巨大的潜力, 现在我们 |
| 58 | using it in real work. | 在真正的工作中使用它。 |
| 59 | It's productive and it's profitable. | 生产而利. |
| 60 | It's generating enormous profits for them. | 它为他们带来了巨大的利润。 |
| 61 | Because the tokens are now profitable and because the token demand is so incredible, | 因为信物现在有利可图 而且信物的需求是如此惊人, |
| 62 | it's the reason why they want more compute because the compute translates to more tokens, | 这就是为什么他们想要更多的计算 因为计算 翻译为更多的符号, |
| 63 | to more profits, and that flywheel for them is going to be great. | 给更多的利润, 而飞轮对他们来说将是巨大的。 |
| 64 | So I think their IPO is just, of course, another day in their journey and it's a very important | 所以,我认为他们的IPO只是,当然, 在他们的旅程的另一天, 这是一个非常重要的 |
| 65 | day for them. | 他们的一天。 |
| 66 | They've raised a lot of money and they can apply that money to fueling their growth. | 他们筹集了很多钱,可以运用这些钱来刺激他们的成长. |
| 67 | So I think that's the two labs. | 我想这是两个实验室。 |
| 68 | Now, we're just looking at the tip of the iceberg with the two of them because they're | 现在,我们只是看着冰山的一角 与他们两人,因为他们 |
| 69 | the most, if you will, visible. | 最,如果你愿意,可见。 |
| 70 | But if you look at the AI industry, it's broad and it's extremely deep. | 但是,如果看AI行业,它的范围很广,而且非常深. |
| 71 | It's also extremely expensive. | 亦极贵. |
| 72 | I mean, when investors see what AI really costs, are they going to be happy or are they | 我的意思是,当投资者看到AI的真正成本, 他们将是快乐还是他们是 |
| 73 | going to decide the numbers don't add up? | 要决定 数字不加起来? |
| 74 | The cost of AI, the cost of a token is reducing by 30, 40 times per year. | 人工智能的费用,代币的费用每年减少30,40倍. |
| 75 | And so I think that the cost of tokens is decreasing. | 所以我认为 代币的成本正在下降。 |
| 76 | The cost of intelligence is also decreasing. | 情报费用也在下降。 |
| 77 | And so the benefits of the customers are actually going up exponentially. | 因此,顾客的利益 正在成倍增长。 |
| 78 | That said, AI is deeply unpopular and people are scared. | 尽管如此,AI却非常不受欢迎,人们也很害怕. |
| 79 | Sam Altman, Dario Amadei, Elon Musk all say we should slow down. | 萨姆·阿尔特曼,达里奥·阿玛迪,埃隆·穆斯克都说我们应该慢下来. |
| 80 | You have been very consistent in your optimism. | 你一直很乐观 |
| 81 | What do you see that they don't and how can they be so wrong? | 你看到什么,他们没有, 他们怎么会这样错? |
| 82 | First of all, safety is paramount and we, in order for us to realize the incredible | 首先,安全是至高无上的 我们,为了让我们认识到 |
| 83 | potentials and the optimistic future that we all hope for and believe in, we have to | 我们所有人都希望并相信的潜力和乐观的未来,我们必须 |
| 84 | develop the technology safely and responsibly. | 安全和负责任地开发技术。 |
| 85 | I believe we have agency in how this technology advances. | 我相信我们有这个技术如何发展的机构。 |
| 86 | We should accelerate the development of open shell. | 我们应该加快开放壳体的发展. |
| 87 | We should accelerate the integration of this containment technology into AI systems. | 我们应该加快将这种封存技术纳入AI系统. |
| 88 | We should accelerate the development of monitoring technologies and AIs that monitor AIs. | 我们应该加速发展监测技术和监测AI的AI. |
| 89 | So we should accelerate AI technology. | 所以我们应该加速AI技术. |
| 90 | And so I think that there's a fallacy. | 所以我认为这是个谬论。 |
| 91 | I think it's in fact a mistake to separate AI capability and AI safety as if it's two | 我觉得把人工智能能力和人工智能安全 分开就像两个 |
| 92 | conversations, it's one conversation. | 谈话,这是一次谈话。 |
| 93 | If one of the labs was doing something that was too dangerous, that you felt was too | 如果其中一个实验室在做一些 太危险的事情,你觉得太 |
| 94 | dangerous, would Nvidia stop selling to them? | 危险,恩维迪亚会停止出售给他们吗? |
| 95 | It's a hypothetical, but of course, yeah, yeah, it's a hypothetical, but of course. | 这是假设,但当然是,当然是,当然是,当然是,当然是假设. |
| 96 | But before you do so, you would have a conversation and, you know, we don't run other | 但在此之前,你会有一个谈话, 你知道,我们不运行其他 |
| 97 | people's companies and there's a limit to the judgment by which we apply to other | 人民公司的判断是有限的 我们适用于其他公司 |
| 98 | companies' use of our technology, our technology's general purpose. | 公司使用我们的技术, 我们的技术的通用目的。 |
| 99 | And so imposing our values on some other people's companies and for example, | 把我们的价值观强加给别人的公司 比如说 |
| 100 | institutions or countries, that's really largely out of bounds. | 机构或国家, 这实际上大多是极限。 |
| 101 | But if we if we know for certain that an application of our technology is bringing harm | 但是如果我们确信 我们技术的应用 正在带来伤害 |
| 102 | to people, of course. | 当然是给人们的 |
该视频共有字幕 102 条。解锁更多字幕为会员功能，请移动到 价格