The taste to know what to build is what matters now.

演讲者指出，在现代产品开发和创新领域，正在发生一场深刻的转变。他认为，如今最重要的技能是那种能够判断“做什么”、 “如何有效地展示这些信息”以及“选择哪种媒介最能达到特定目标”的“品味”或独到眼光。这一转变主要源于实施（或执行）前所未有的简便性和经济性。 得益于特别是人工智能和大型语言模型等方面的进步，技术准入门槛已大幅降低。正如演讲者所说，“任何人都能构建任何东西”，并且可以“以惊人的速度和极低的成本迅速实现你想要的任何功能”。这种唾手可得的技术解决方案的普及，使得“在各种媒介上的实施都变得如此廉价”。 这种便捷性带来的直接后果是，人们强烈倾向于绕过传统的、往往费时费力的规划阶段。演讲者指出，一种普遍的情绪是团队“非常想直接跳到原型阶段”，常常宣称“PRD已死”，转而倾向于立即展示概念。在这种新环境下，快速展示“我所指为何”的魅力，远胜过细致地记录下来。 然而，演讲者也指出了工程师中一个截然不同却同样成问题的倾向：即乐于“撰写大量文档”，但这些文档最终却被证明“不值得一读”。这表明，尽管构建或撰写文档的“行为”可能很容易，但产出有“价值”的成果——无论是代码还是沟通内容——仍然是一项独特的挑战。仅仅生成更多内容并不等同于有效的沟通或战略上的协调一致。 因此，核心观点是，正因为实施变得如此普遍和廉价，重心必须转向战略性沟通。选择“适合你想要表达观点的正确形式”变得“非常重要”。这需要一个深思熟虑的“策展过程”，其中需要仔细考虑选择最具影响力和最合适的媒介来传达想法，无论是简洁的原型、集中的设计，还是结构良好且相关的文档。最终的成功不再取决于构建的能力，而在于选择、阐明和有效指导该构建过程的智慧。

The speaker identifies a profound shift in the modern landscape of product development and innovation, asserting that the most critical skill is now the "taste" or discerning judgment required to determine *what* to work on, *how* to effectively present that information, and *which medium* is best suited to achieve specific goals. This transformation is primarily driven by the unprecedented ease and affordability of implementation. Thanks to advancements, particularly in AI and Large Language Models, the technical barriers to entry have dramatically lowered. As the speaker notes, "anybody can build anything," and one can "stand up whatever feature you want" with surprising speed and minimal cost. This abundance of readily available technical solutions has made implementation "so cheap across every medium." A direct consequence of this ease is the strong temptation to bypass traditional, often laborious, planning stages. The speaker highlights a common sentiment where teams are "very tempting to jump straight to a prototype," often proclaiming that "PRDs are dead" in favor of immediately demonstrating a concept. The allure of quickly showing "what I mean" rather than meticulously documenting it is powerful in this new environment. However, the speaker also points out a contrasting, yet equally problematic, tendency among engineers: the inclination to "write a lot of documents" that ultimately prove "not worth reading." This suggests that while the *act* of building or documenting might be easy, producing *valuable* output – whether it's code or communication – remains a distinct challenge. Simply generating more content doesn't equate to effective communication or strategic alignment. Therefore, the core insight is that precisely because implementation has become so abundant and inexpensive, the emphasis must shift towards strategic communication. It becomes "really important to pick the right format for the point you're trying to make." This necessitates a deliberate "curation process" where careful consideration is given to selecting the most impactful and appropriate medium for conveying ideas, whether that's a concise prototype, a focused design, or a well-structured, relevant document. The ultimate success hinges not on the capacity to build, but on the wisdom to choose, articulate, and guide that building process effectively.