Stanford Graduate School of Business - Jensen Huang, Founder and CEO of NVIDIA
发布时间:2024-03-06 00:13:14
原节目
黄仁勋重返斯坦福,进行了一场引人入胜的对话,内容围绕着他作为创始人的历程以及英伟达的演变展开。他首先回忆了自己离开LSI Logic的情形,他受到朋友克里斯和柯蒂斯的激励,渴望解决通用计算机无法解决的问题。
黄仁勋强调了拥有强大的“过往”并展示基础技能的重要性。他提到了自己早期在丹尼餐厅洗碗的经历,在那里他学习了组织和效率,为他未来作为CEO的成功奠定了基础。他幽默地否定了传统的商业计划,选择依靠自己的声誉和有说服力的沟通技巧。他回忆起自己在LSI Logic的前任CEO威尔·科里根是如何安排他与红杉资本的创始人唐·瓦伦丁会面的。
对话深入探讨了英伟达的早期,以及它对视频游戏3D图形的关注,而当时这个市场还不存在。黄仁勋回忆起他所面临的怀疑,以及有人提醒他,他所依靠的是电子艺界的一位14岁少年,并且是由他的母亲驱动的。他强调了英伟达的双重使命,即创造技术和市场,例如它早期进入自动驾驶、深度学习和计算药物设计领域。
一个关键时刻是英伟达决定采用OpenGL标准,这是微软宣布Direct3D之后的一个转折点,重塑了公司,并灌输了一种即使在没有事先知识的情况下也能创新的信念。黄仁勋强调了在情况变化时重新审视第一性原理并重新发明解决方案的重要性。这种心态促使他们在可编程着色器方面进行了开创性的工作,并开发了CG,这是CUDA的前身,CUDA是GPU的语言。
黄仁勋强调了“未来成功的早期指标”(EOIFS)的重要性,而不是传统的KPI,他强调工作的意义,而不是眼前的经济回报。这种方法指导了英伟达对深度学习的投资,即使当时市场不存在,也是因为他们相信这是一项“有价值的工作”。他讨论了该公司愿意承担没有直接经济回报的项目。
在讨论挑战时,黄仁勋回忆了英伟达在金融危机期间市值缩水80%的情况,强调了回归核心信念并保持专注的必要性。他将自己的领导风格描述为高度参与,直接向他汇报的有50人,沟通开放,并致力于向人们展示如何推理复杂的难题。他提倡扁平化的组织结构,信息自由流动,使员工能够做出明智的决策,并培养透明和信任的文化。
关于人工智能的话题,黄仁勋强调了生成式人工智能的变革潜力,它使计算机能够理解和翻译不同数据模式。他认为,这种转变将导致一个更具生成性的计算未来,影响从网络和存储到软件开发等各个行业。
黄仁勋还分享了他对人工智能监管的看法,主张现有监管机构调整其框架,以解决人工智能特有的问题,而不是制定可能扼杀发展的总体监管。他强调了在接地、安全和网络安全等领域的技术进步对于减轻与人工智能相关的风险的重要性。
在会议结束时,黄仁勋强调了为世界做出独特贡献、过上有目标的生活以及围绕在自己周围支持自己的人的重要性。他鼓励听众拥有一个核心信念,每天反思,并全力以赴地追求梦想。他表达了英伟达对推进计算未来的承诺。
Jensen Huang's return to Stanford was marked by an engaging conversation about his journey as a founder and the evolution of NVIDIA. He began by recounting his departure from LSI Logic, driven by a desire to solve problems beyond the capabilities of general-purpose computers, spurred by his friends Chris and Curtis.
Huang emphasized the importance of having a strong "past" and demonstrating foundational skills, referencing his early experience as a dishwasher at Denny's where he learned organization and efficiency, laying the groundwork for his future success as a CEO. He humorously dismissed traditional business plans, opting to rely on his reputation and persuasive communication skills. He recounted how his former CEO at LSI Logic, Will Corrigan, got him a meeting with Don Valentine, the founder of Sequoia Capital.
The conversation delved into the early days of NVIDIA and its focus on 3D graphics for video games, a market that didn't yet exist. Huang recounted the skepticism he faced, and the reminder that a 14-year-old at Electronic Arts, driven by his mom, was who he was relying on. He highlighted NVIDIA's dual mission of creating both technology and markets, exemplified by its early entry into autonomous driving, deep learning, and computational drug design.
A pivotal moment was NVIDIA's decision to adopt the OpenGL standard, which was a turning point after Direct3D was announced by Microsoft, reshaping the company and instilling a belief that they could innovate even without prior knowledge. Huang emphasized the importance of revisiting first principles and reinventing solutions in light of changing conditions. This mindset led to their pioneering work in programmable shaders and the development of CG, a precursor to CUDA, which was the language for GPUs.
Huang stressed the importance of "Early Indicators Of Future Success" (EOIFS) over traditional KPIs, focusing on the importance of the work rather than immediate financial returns. This approach guided NVIDIA's investment in deep learning, even when the market was nonexistent, driven by the belief that it was "worthy work." He discussed the company's willingness to take on projects with no immediate financial returns.
Discussing challenges, Huang recounted NVIDIA's 80% market cap loss during the financial crisis, emphasizing the need to return to core beliefs and maintain focus. He described his leadership style as highly engaged, with 50 direct reports, open communication, and a commitment to showing people how to reason through complex problems. He advocates for a flat organizational structure where information flows freely, empowering employees to make informed decisions, and fostering a culture of transparency and trust.
On the topic of AI, Huang highlighted the transformative potential of generative AI, enabling computers to understand and translate between different modalities of data. This shift, he argued, will lead to a more generative future of computing, impacting industries from networking and storage to software development.
Huang also shared his perspective on AI regulation, advocating for existing regulatory bodies to adapt their frameworks to address AI-specific concerns, rather than creating overarching, potentially stifling regulations. He stressed the importance of technological advancements in areas like grounding, safety, and cyber security to mitigate the risks associated with AI.
Concluding the session, Huang emphasized the importance of making a unique contribution to the world, living a life of purpose, and surrounding oneself with supportive individuals. He encouraged the audience to have a core belief, to check in everyday, and to pursue the dream with all their might. He expressed NVIDIA's commitment to advancing the future of computing.