Zynga 创始人马克·平卡斯分享他的产品行动指南，以及从网络、社交和移动革命中汲取的 AI 时代构建产品的见解。
Mark Pincus, founder of Zynga, shares his product playbook and insights on building for the AI era, drawing from his experiences across the web, social, and mobile revolutions.
要点 · TL;DR
消费级 AI 现在不可投资，但提供了最大的机会来创造新的互联网宝藏。 Consumer AI is uninvestable now but offers the greatest opportunity for new internet treasures.
使用“已验证、更好、新”框架：复制已验证的，改进更好的，测试新想法并预期失败。 Use Proven Better New framework: copy proven, improve better, test new ideas expecting failure.
创始人模式意味着相信直觉并为转型创造环境，而不仅仅适用于精英创始人。 Founder mode means trusting instincts and creating context for pivots, not just for elite founders.
核心观点 · Key points
尽管消费者 AI 目前不可投资，但机会空前巨大，有望诞生新的互联网宝藏。 Consumer AI opportunity is greatest now despite being uninvestable, with potential for new internet treasures.
Proven Better New 框架：复制已证明的，改进更好的，测试新想法并假设它们会失败。 Proven Better New framework: copy proven, improve better, test new ideas assuming they'll fail.
创始人模式意味着允许自己跟随直觉，并为转型创造环境。 Founder mode means giving yourself permission to follow instincts and create context for pivots.
管理的目标是让员工在你不在场时做正确的事。 Management's goal is to get people to do the right thing when you're not in the room.
真正的产品热度（鱼群涌动）是毋庸置疑的；创始人应收集胜利，而非下注。 True product heat (fish running) is unmistakable; founders should collect winnings, not make bets.
免费是强大的模式；在 AI 时代，免费无限算力将重塑消费者服务。 Free is a powerful model; in AI era, free unlimited compute will reshape consumer services.
反共识 · Contrarian takes
消费者 AI 现在不可投资，但机会空前巨大——与风投聚焦企业级相反。 Consumer AI is uninvestable now, yet the opportunity has never been greater—contrary to VC focus on enterprise.
大多数企业尚未从 AI 获益；这是技能问题，而非技术失败。 Most enterprises see no AI benefit yet; it's a skill issue, not a technology failure.
不要编写调用 LLM 的代码；应编写教 LLM 写代码的标记语言。 Don't write code that calls LLMs; write markdown that teaches LLMs to write code.
创始人模式适用于每一位创始人，而非仅限贝索斯等精英——与风投观点相反。 Founder mode is for every founder, not just elite like Bezos—contrary to VC opinion.
消费者 AI 革命可能在 2029 年，距离今天还有三个数量级。 The consumer AI revolution is likely in 2029, three orders of magnitude away from today.
新想法（Proven Better New 中的新）几乎总是失败；这是预料之中的，不应气馁。 New ideas (new in Proven Better New) almost always fail; that's expected and should not demoralize.
本期章节 · Chapters（共 11）
引言与写作动机Introduction and Book Motivation
全天候 AI 助手概念Always-on AI assistant concept
消费者分发挑战Consumer distribution challenges
成熟、更优、新框架Proven, better, new framework
管理团队士气与转型Managing team morale and pivots
创始人模式与领导力Founder Mode and Leadership
AI 影响时间线与代币消耗Timeline for AI impact and token spending
消费者服务与未来机遇Consumer services and future opportunities