Netflix CPTO on AI and the future of product and tech roles | Elizabeth Stone

以下是内容的中文翻译： Netflix 产品与技术主管伊丽莎白·斯通 (Elizabeth Stone) 强调了人工智能，特别是生成式人工智能，对传统技术岗位的变革性影响。她观察到，目前正处于一个“风暴期”（storming phase），产品经理在写代码，设计师在撰写产品需求文档（PRD），而工程师则深入研究产品策略，这自然导致了职责的混乱。她认为这种职责的灵活性可以加速开发，让产品和设计团队更快地进行原型设计，但她也警告不要让每个人都“做所有事情”。在工程、数据科学和设计等专业职能中，精益求精的专业技能仍然至关重要且稀缺。 Netflix 正在通过强调“系统性思考者”来适应这种变化。对于工程部门，这意味着从本地化的团队特定技术栈转向通用基础设施和核心能力的“预设路径”（paved paths），这对于管理跨多个系统的 AI 代理至关重要。设计师们越来越专注于开发系统级的模板和设计语言，以确保产品之间的一致性，并让非设计师也能在品牌指南下进行构建。这需要一种全新的思维模式——一种能够适应模糊性、敢于承担风险并持续探索的思维模式，这既适用于现有员工，也适用于新招聘的人才。 斯通指出，专业分工正在从非常狭窄的专业化转向更具适应性的通才。虽然在某些领域（例如编码、播放系统）仍然需要利基专业知识，但重点在于那些能够快速学习并在不同领域应用技能，从而掌握“更广泛工具集”的个体。对于有抱负的系统性思考者，她的建议是“跳出一步”，从眼前的问题中抽离出来，质疑更广泛的假设，并考虑经理或同事的视角，以理解个人贡献如何融入更大的组织生态系统。 在职业发展和招聘方面，Netflix 在从初级到高级的所有岗位上都提倡“AI 流利度”。这不一定意味着亲自动手构建，而是指具备关于何时以及如何有效利用 AI 的良好判断力，同时对新技术保持开放态度。Netflix 继续招聘初级人才，重视他们对 AI 的天生熟悉度以及对不断变化的娱乐和消费者行为的全新视角。虽然 AI 工具简化了一些任务，但产品质量、设计、测试以及理解“代码、计算机系统和产品如何运作”的核心责任仍然至关重要。斯通强调，工程领域将逐渐适应 AI 代理编写代码，但工程师仍需理解底层系统，以确保质量、调试和指导技术发展。 除了编码和原型设计，AI 还深刻影响着 Netflix 的运营。它增强了数据分析能力，从实验、消费者洞察和业务利益相关者那里提炼出大量信息。这有助于更快地生成假设和可操作的洞察。在内容制作方面，AI/ML 长期以来一直被用于视觉效果和本地化。现在，生成式 AI 扩展到创意构思（预可视化）、后期制作（例如，通过获取中间正片进行重新打光、重新构图、对白修改）以及在全球范围内扩展推广素材、艺术作品、字幕和配音。Netflix 过去在 AI/ML 方面的投入，例如通过“Netflix 大奖”来推动个性化，使其在利用这些技术应对日益增长的内容广度和多样化格式方面占据先机。 Netflix 的企业文化，以“卓越即操作系统”为特色，强调高人才密度、自主性和责任感。这一理念，与当今顶级 AI 实验室的运作方式不谋而合，意味着赋予有才华的个体以自主权，将决策权下放至组织深处，并信任他们的判断力。关键要素包括：将人才密度作为不可妥协的基础；乐于承担风险并从失败中学习；无私地专注于“Netflix 至关重要，Netflix 会员至关重要”；以及抵制在出现问题时增加流程的冲动。“留用测试”（keepers test）——即持续评估今天是否会重新雇用某位员工——既是积极反馈的机制，也是进行艰难对话的工具，确保人才标准始终保持高位。 展望未来，斯通预见娱乐将超越传统的电影和电视，涵盖移动/云游戏、直播内容和播客，提供更丰富的多样性和沉浸感。Netflix 旨在简化这些多样化格式的发现和互动，使体验更加个性化。尽管 AI 将为创作者提供新的工具和可能性，但斯通坚信，人类仍将是故事叙述的核心，提供定义娱乐的必要连接和情感。

Elizabeth Stone, Netflix's Product and Technology Officer, highlights the transformative impact of AI, particularly Generative AI, on traditional tech roles. She observes a "storming phase" where PMs ship code, designers write PRDs, and engineers delve into product strategy, leading to a natural confusion about job responsibilities. While she believes this fluidity can accelerate development, allowing product and design to prototype faster, she cautions against everyone "doing everything." Craft excellence in specialized functions like engineering, data science, and design remains crucial and scarce. Netflix is adapting by emphasizing "systems thinkers." For engineering, this means moving from localized team-specific stacks to common infrastructure and "paved paths" for core capabilities, essential for managing AI agents across multiple systems. Designers are increasingly focused on developing system-wide templates and design languages to ensure coherence across products, enabling non-designers to build within brand guidelines. This requires a new mindset—one comfortable with ambiguity, risk-taking, and continuous exploration, both for existing talent and new hires. Stone notes a shift away from very narrow specializations towards more adaptable generalists. While niche expertise is still needed in certain areas (e.g., encoding, playback systems), the emphasis is on individuals who can quickly learn and apply their skills across different domains, fostering a "broader array of tools." For aspiring systems thinkers, her advice is to "step out one click" from the immediate problem and question broader assumptions, considering the manager's or colleagues' perspectives to understand how individual contributions fit into the larger organizational ecosystem. In terms of career development and hiring, Netflix promotes "AI fluency" across all roles, from junior to senior. This doesn't necessarily mean hands-on building, but rather having good judgment about when and how to use AI effectively, alongside an open-minded approach to new technologies. Netflix continues to hire junior talent, valuing their native familiarity with AI and fresh perspectives on evolving entertainment and consumer behaviors. While AI tools simplify some tasks, the core responsibility for product quality, design, testing, and understanding "how code, computer systems, and products work" remains paramount. Stone stresses that engineering will evolve to be comfortable with agents writing code, but engineers will still need to understand the underlying systems to ensure quality, debug, and guide the technology. Beyond coding and prototyping, AI profoundly impacts Netflix's operations. It enhances data analysis, distilling vast amounts of information from experiments, consumer insights, and business stakeholders. This allows for faster hypothesis generation and actionable insights. In content production, AI/ML have long been used in visual effects and localization. Now, Generative AI extends to creative ideation (pre-visualization), post-production (e.g., Interpositive acquisition for re-lighting, reframing, dialogue changes), and scaling promotional assets, artwork, subtitles, and dubs globally. Netflix's historical investment in AI/ML, exemplified by the "Netflix Prize" for personalization, gives it a running start in leveraging these technologies to address the increasing breadth of its content and diverse formats. Netflix's culture, characterized by "excellence as an operating system," emphasizes high talent density, autonomy, and accountability. This philosophy, presciently aligned with how top AI labs operate today, means giving talented individuals agency, pushing decision-making deep into the organization, and trusting their judgment. Key ingredients include: talent density as a non-negotiable foundation; comfort with risk-taking and learning from failures; a selfless focus on "Netflix matters, Netflix members matter"; and resisting the urge to add process when things go wrong. The "keepers test"—continuously evaluating whether one would re-hire an employee today—serves as a mechanism for both positive feedback and difficult conversations, ensuring the bar for talent remains high. Looking ahead, Stone envisions entertainment evolving beyond traditional film and TV to include mobile/cloud games, live content, and podcasts, offering greater variety and immersion. Netflix aims to simplify discovery and engagement across these diverse formats, making the experience more personalized. While AI will enable creators with new tools and possibilities, Stone firmly believes humans will remain at the heart of storytelling, providing the essential connection and emotion that defines entertainment.