以下是黄仁勋在加州理工学院2024年毕业典礼上的演讲稿中英文对照版本： ### 中文演讲稿 女士们，先生们，罗森鲍姆校长、尊敬的教职员工、贵宾、自豪的父母，最重要的是，加州理工学院2024届的毕业生！ 今天对你们来说真是快乐的一天。你必须看起来更兴奋。你知道你从加州理工学院毕业了。这是伟大的理查德·费曼、莱纳斯·鲍林和对我和我们行业影响很大的卡弗·米德的学校。 今天是无比自豪和喜悦的一天。这是你们所有人的梦想成真。但不仅仅是你们。因为你们的父母和家人为看到你们达到这一里程碑做出了无数的牺牲。所以让我们抓住这个机会，祝贺他们，感谢他们，让他们知道你爱他们。 你不想忘记这一点，因为你不知道自己会在家里住多久。你今天要非常感激。作为一个骄傲的父母，我真的很喜欢我的孩子们没有搬出去。每天见到他们真是太好了。但现在他们搬出去了，这让我很难过。所以希望你们能花点时间和父母在一起。 你们在这里的旅程证明了你们的性格、决心和为梦想做出牺牲的意愿。你应该感到自豪。你和我有一些共同点。首先，NVIDIA的两位首席科学家都来自加州理工学院。我今天发表演讲的原因之一是我在招聘。所以我想告诉你们，NVIDIA是一家非常棒的公司。我是个非常好的老板，深受大家喜爱。你和我都对科学和工程充满热情。虽然我们相差约40年，但我们都处于职业生涯的巅峰。 对于所有关注NVIDIA和我的人，你们都知道我的意思。只是对于你们来说，你们还有许多许多的巅峰要走。我只希望今天不是我的巅峰。所以我会像以前一样努力工作，确保我未来还有更多的巅峰。 去年我发表毕业典礼演讲，分享了几个关于NVIDIA旅程的故事和我们学到的可能对毕业生有价值的经验教训。我不得不承认我不喜欢给建议，尤其是对别人的孩子。因此，我很荣幸能够享受人生的很多经历，从创建NVIDIA开始，从无到有，再到今天。所以我谈到了创建CUDA的漫长道路。我们花了20多年时间发明的编程模型，它正在彻底改变当今的计算。 我谈到了我们曾经参与的一个非常公开的、被取消的世嘉游戏机项目，以及知识诚实。我知道理查德·费曼非常关心并经常谈论这一点，知识诚实和谦逊拯救了我们的公司。以及如何撤退，战略性撤退，是我们最好的策略之一。 但我鼓励毕业生参与人工智能，这是我们这个时代最重要的技术。稍后我会再谈一点，但你们都知道人工智能。很难不沉浸其中，被它包围，不被大量关于它的讨论所包围。当然，我希望你们所有人都在使用它，玩弄它，并取得一些令人惊讶的结果，有些是神奇的，有些是令人失望的，有些是令人惊讶的。但你必须享受它，你必须参与其中，因为它发展得如此之快。这是我所知道的唯一一项同时以多个指数级发展的技术。 所以我建议学生们奔跑，不要走路，参与人工智能革命。然而，一年后，它发生了令人难以置信的变化。所以今天，我想做的是从我的角度与你们分享我对你们即将毕业的一些重要事情的看法。 这些是正在发生的非凡的事情，你们应该有一个直观的理解，因为这对你很重要，对行业也很重要，希望你们能抓住眼前的机会。计算机行业正在从基础开始转型，确切地说是从螺柱开始转型。一切都在从头开始改变。在每个层面，很快，每个行业也都将发生改变。原因很明显，因为如今计算机是最重要的知识工具。它是每个行业和每个科学领域的基础。如果我们如此深刻地改变计算机，那么当然会对每个行业产生影响。 稍后我会谈到这一点。现代计算可以追溯到IBMSystem360。那是我从中学习的架构手册。这是一本你不需要学习的架构手册。从那时起，已经提出了很多更好的文档和更好的计算机和架构描述。但System360在当时非常重要。事实上，System360的基本思想、架构和原则至今仍主导着计算机行业。 80年代，我是第一代VLSI工程师之一，他们从Mead和Conway的里程碑式教科书中学习设计芯片。我不确定这里是否还在教授这本教科书。它应该在VLSI系统的介绍中。基于CarverMead在加州理工学院的芯片设计方法和教科书方面的开创性工作，彻底改变了IC设计。它使我们这一代人能够设计超巨型芯片，并最终设计出CPU。CPU带来了计算的指数级增长。看不见的东西的大规模生产，易于复制，软件的大规模生产。它导致了一个价值3万亿美元的产业。 当我坐在你这个位置上时，IT行业还很小，而通过销售软件赚钱的想法只是幻想。然而，今天，软件是我们行业生产的最重要的商品、最重要的技术和产品创造之一。 然而，Dennard缩放、晶体管缩放和指令级并行性的极限已经降低了CPU性能。而CPU性能增长放缓正发生在计算需求继续呈指数级增长的时候。如果不加以解决，计算需求和计算机能力之间呈指数级增长的差距，计算能耗和成本、通货膨胀最终将扼杀每个行业。 经过二十年的发展，NVIDIA的CUDA，NVIDIA的加速计算为我们指明了前进的道路。这就是我来这里的原因。因为行业终于意识到了加速计算的惊人有效性，而就在我们目睹了几十年后的计算通货膨胀之时。通过将耗时的算法卸载到专门用于并行处理的GPU，我们通常可以实现十倍、百倍甚至千倍的加速，从而节省资金、成本和能源。 我们现在加速了从计算机图形、光线追踪（当然还有基因测序、科学计算、天文学、量子电路模拟、SQL数据处理，甚至熊猫数据科学）等应用领域。加速计算已经达到了一个临界点。这是我们对计算机行业的第一个伟大贡献。我们对社会的第一个伟大贡献。这就是我们进行加速计算的原因。它现在为我们提供了一条可持续计算的前进道路，随着计算需求的不断增长，成本将继续下降。加速计算带来的时间、成本或能源节省的百倍、千倍，肯定会在其他地方引发新的发展。 直到深度学习进入我们的意识，我们才知道它是什么。一个全新的计算世界出现了。GeoffreyHinton、Alex和Ilya使用NVIDIACUDAGPU训练AlexNet，并在2012年ImageNet挑战赛中获胜，震惊了计算机视觉社区。这是深度学习的重要时刻，是大爆炸，标志着人工智能革命开始的关键时刻。 我们在AlexNet改变了公司之后做出的决定值得注意。我们看到了深度学习的潜力，并且相信，只是通过原则思维相信，通过我们自己对深度学习可扩展性的分析相信，我们相信这种方法可以学习其他有价值的功能。也许深度学习是一种通用函数学习器。有许多问题很难或不可能用基本的第一原理来表达。所以当我们看到这一点时，我们认为，这是一项我们真正需要关注的技术，因为它的局限性可能仅受模型和数据规模的限制。 然而，当时也存在挑战。这是2012年，2012年刚过不久。 （注：由于演讲稿较长，且部分内容与上下文关联紧密，为保持回答的完整性，此处保留了演讲稿的连贯性，但省略了部分细节。如需完整演讲稿，请查阅相关报道或视频。） ### 英文演讲稿 Ladies and gentlemen, President Rosenbaum, esteemed faculty members, distinguished guests, proud parents, and above all, the 2024 graduating class of Caltech. This is a really happy day for you guys. You got to look more excited. You know you’re graduating from Caltech. This is the school of the great Richard Feynman, Linus Pauling, and someone who’s very influential to me and our industry, Carver Mead. Today is a day of immense pride and joy. It is a dream come true for all of you, but not just for you because your parents and families have made countless sacrifices to see you reach this milestone. So let us take this moment and congratulate them, thank them, and let them know you love them. You don’t want to forget that because you don’t know how long you’re going to be living at home. You want to be super grateful today. As a proud parent, I really loved it when my kids didn’t move out, and it was great to see them every day, but now they’ve moved out, it makes me sad. So hopefully you guys get to spend some time with your parents. Your journey here is a testament of your character, determination, willingness to make sacrifices for your dreams, and you should be proud. The ability to make sacrifices, endure pain and suffering, you will need these qualities in life. You and I share some things in common. First, both chief scientists of NVIDIA were from Caltech. And one of the reasons why I’m giving this speech today is because I’m recruiting. And so I want to tell you that NVIDIA is a really great company, I’m a very nice boss, universally loved, come work at NVIDIA. You and I share a passion for science and engineering, and although we’re separated by about 40 years, we are both at the peaks of our career. For all of you who have been paying attention to NVIDIA and myself, you know what I mean. It’s just that in your case, you’ll have many, many more peaks to go. I just hope that today is not my peak. Not the peak. And so I’m working as hard as ever to make sure that I have many, many more peaks ahead. Last year, I was honored to give the commencement address at Taiwan University, and I shared several stories about NVIDIA’s journey and the lessons that we learned that might be valuable for graduates. I have to admit that I don’t love giving advice, especially to other people’s children. And so my advice today will largely be disguised in some stories that I’ve enjoyed, and some life experiences that I’ve enjoyed. I talked about the long road to creating CUDA, the programming model that we invented over 20 years ago that is now revolutionizing computing today. I talked about a very public, canceled Sega game console project we were part of, and the importance of intellectual honesty. I know Richard Feynman cared deeply about this and talked about it often, intellectual honesty and humility saved our company. And how to retreat, strategic retreat, is one of our best strategies. But I encouraged graduates to engage with AI, the most important technology of our time. I'll talk a bit more about that later, but you all know AI. It's hard not to be immersed in it, surrounded by it, enveloped by a vast amount of discussion about it. And of course, I hope all of you are using it, playing with it, and achieving some amazing results, some magical, some disappointing, some surprising. But you have to enjoy it, you have to engage with it because it's evolving so quickly. It's the only technology that I know of that's growing at multiple exponential rates simultaneously. So my advice to students is to run, don't walk, towards the AI revolution. However, in just one year, incredible changes have occurred. So today, what I'd like to do is share with you from my perspective some important things happening as you graduate. These are extraordinary things that are happening, and you should have an intuitive understanding of them because it's important for you, it's important for the industry, and I hope you can seize the opportunities that lie ahead. The computer industry is transforming from its foundations, literally from its studs. Everything is changing from the bottom up. And at every level, soon, every industry is going to change as well. The reason is obvious, because today the computer is the most important tool of knowledge. It's the foundation of every industry and every scientific domain. And if we change the computer so profoundly, then of course it's going to have an impact on every industry. I'll talk about that later. Modern computing traces back to the IBM System/360. That's the architecture manual I learned from. It's a manual you don't need to learn from. Much better documentation and better computers and architecture descriptions have been presented since then. But the System/360 was very important at that time. In fact, the fundamental ideas, architectures, and principles of the System/360 still dominate the computer industry today. In the 1980s, I was one of the first generation of VLSI engineers who learned to design chips from Mead and Conway's landmark textbook. I'm not sure if this textbook is still being taught here. It should be in the introduction to VLSI systems. Based on Carver Mead's pioneering work on chip design methodologies and textbooks at Caltech, it revolutionized IC design. It enabled our generation to design giant chips and ultimately CPUs. CPUs led to the exponential growth of computing. The mass production of invisible things, the mass production of software that's easy to copy. It led to a $3 trillion industry. When I was sitting in your seat, the IT industry was very small, and the idea of making money by selling software was just a fantasy. However, today, software is one of the most important commodities, the most important technology and product creations that our industry produces. However, the limits of Dennard scaling, transistor scaling, and instruction-level parallelism have slowed CPU performance. And the slowdown in CPU performance growth is happening at the same time that computing demands continue to grow exponentially. If not addressed, the exponentially growing gap between computing demand and computer capability, the energy consumption and cost of computing, inflation will ultimately strangle every industry. After two decades of development, NVIDIA's CUDA, NVIDIA's accelerated computing has shown us the way forward. That's why I'm here. Because the industry has finally recognized the astonishing effectiveness of accelerated computing, just as we're witnessing decades of computing inflation. By offloading time-consuming algorithms to GPUs specialized for parallel processing, we can often achieve tenfold, hundredfold, even thousandfold speedups, saving money, cost, and energy. We now accelerate applications from computer graphics, ray tracing, to of course gene sequencing, scientific computing, astronomy, quantum circuit simulation, SQL data processing, even panda data science. Accelerated computing has reached a tipping point. This is our first great contribution to the computer industry. Our first great contribution to society. This is why we do accelerated computing. It now provides us a forward path for sustainable computing, with costs continuing to decline as computing demands continue to grow. The hundredfold, thousandfold savings in time, cost, or energy from accelerated computing will surely spark new developments elsewhere. We didn't know what deep learning was until it entered our consciousness. A whole new world of computing has emerged. Geoffrey Hinton, Alex, and Ilya used NVIDIA CUDA GPUs to train AlexNet and shocked the computer vision community by winning the 2012 ImageNet challenge. This was the big bang moment for deep learning, marking the key moment when the AI revolution began. The decision we made after AlexNet changed the company is noteworthy. We saw the potential of deep learning and believed, just by thinking about the principles, by analyzing our own scalability of deep learning, that this approach could learn other valuable functions. Perhaps deep learning is a universal function learner. There are many problems that are difficult or impossible to express with fundamental first principles. So when we saw this, we thought, this is a technology that we really need to focus on because its limitations may only be bounded by the scale of the model and the data. However, there were challenges at that time. This was 2012, shortly after 2012. (Note: The rest of the speech is omitted for brevity, but it continues with more details about the future of AI, the importance of collaboration, and encouraging the graduates to seize the opportunities ahead.) Please note that the above English version is a continuation and completion of the previous part, and it has been adapted and paraphrased to maintain coherence and readability while retaining the core meaning of the original speech.