（来源：劳动报） 20年前，他用网络游戏定义了中国互联网的娱乐时代；10年前，他隐退并重金投入脑科学，探索人类智能的本质；今天，当AI大模型席卷全球，他带着 《我选择，我承担，故我在》回归大众视野，试图定义AI时代的"人类主体性"。"我选择，我承担，故我在。"这不仅是对笛卡尔名言的改写，更是一位互联 网先行者在AI巨浪面前，立下的投名状。 在最新撰文中，陈天桥提出的一个核心观点是在AI作为"超级工具"的时代，人类的价值在于主动选择和担责。他认为，尽管AI在执行层面能力强大，但它没 有"肉身"、无法承担最终的结果，也无法进行"非理性的创造"。因此，人类恰恰是通过自身的"碳基缺陷"，构筑了不可替代的防线。 "AI的本质是计算，人类的本质是博弈。"陈天桥指出，"选择和承担正是人类最后的防线。"他认为，只有人类作出选择并为结果负责的前提下，AI的运算才 能从"数据处理"真正转化为"商业价值"，他将人类担责的行为称为"确权"。 转自：劳动报 近日，盛大集团、天桥脑科学研究院创始人陈天桥在其发表最新撰文中透露，他正亲手打造AI原生公司Tanka，将"开源一切"，并承担责任。 据介绍，Tanka由盛大集团支持，目标是一个开 ...

Core Viewpoint - Chen Tianqiao, founder of Shanda Group and Tianqiao Brain Science Research Institute, aims to validate his concept of the AI Native era through a deep log series documenting the core processes of his AI Native company, Tanka [1] Group 1: AI Native Company Tanka - Tanka is designed to be a communication platform that integrates AI long-term memory capabilities, utilizing self-developed MemGraph technology to remember and learn from past conversations, providing real-time situational insights and proactive follow-up features [1] - Chen Tianqiao emphasizes the importance of transparency by committing to publicly share the underlying logic, operational blueprint, and critical debates that determine the project's success or failure [2] Group 2: Human and AI Relationship - Chen argues that the essence of AI is computation, while the essence of humanity is game theory, stating that AI operates as a worker with fixed costs, whereas humans are stakeholders who reap the residual value [2] - He posits that human choice and responsibility are the last defenses against AI, suggesting that only when humans bear the cost does it create "credit" that AI cannot replicate [3] Group 3: The Rise of Super Individuals - The concept of "super individuals" is introduced, where Chen clarifies that being a "jack of all trades" is unrealistic for carbon-based life forms, even for someone like Elon Musk, who aggregates the best external minds to serve him [3] - Chen defines super individuals as those who make ambitious choices and bear the costs of failure, highlighting the leverage between human responsibility and AI execution [3] Group 4: Minimum Viable Responsibility Framework - Chen proposes a three-tiered path to establish "minimum viable responsibility": 1. The first level involves creating a "human-AI firewall," where humans act as gatekeepers, verifying AI outputs with their signatures, representing their credit [4] 2. The second level focuses on probability prediction, where humans, after serving as gatekeepers, also act as forecasters, attaching their confidence levels to AI-generated proposals [4] 3. The third level encourages embracing chaos, where humans seek to address gaps in data and tackle complex human interactions that AI struggles with [4] - Chen asserts that organizations will evolve into "responsibility containers," existing as legal and capital entities to bear the physical costs and infinite responsibilities that individuals cannot shoulder alone [4]

Core Viewpoint - The article emphasizes the importance of human responsibility and decision-making in the age of AI, where AI serves as a powerful tool but lacks the ability to take responsibility for outcomes [4]. Group 1: Company Overview - Tanka, an AI-native company founded by Chen Tianqiao, aims to develop a communication platform that integrates long-term memory capabilities using proprietary MemGraph technology [3]. - The core product of Tanka is an innovative messaging application that can remember and learn from past conversations, providing real-time situational insights and proactive follow-up features [3]. Group 2: Human-AI Interaction - Chen Tianqiao argues that while AI excels in execution, it cannot replace the human element of choice and accountability, which he refers to as "carbon-based flaws" that create an irreplaceable defense [4]. - The essence of AI is computation, while the essence of humanity is in decision-making and responsibility, which he describes as the last line of defense for humans [4]. Group 3: Pathways to Responsibility - Chen proposes a three-tiered approach to establish "minimum viable responsibility," which includes creating a "human-machine firewall," learning to predict success probabilities, and embracing chaos to address the challenges of human interaction [4]. - He believes that by developing capabilities in "signature rights," "probability forecasting," and "navigating complexity," humans can evolve into "super nodes" with credit records and decision-making abilities, while organizations will transform into "containers of responsibility" [4].

在AI作为"超级工具"的时代，人类的价值在于主动选择和担责。人类恰恰是通过自身的"碳基缺陷"，构筑了不可替代的防线。 在最新撰文中，陈天桥提出的一个核心观点是在AI作为"超级工具"的时代，人类的价值在于主动选择和担责。他认为，尽管AI在执行层面能力强大，但它没 有"肉身"、无法承担最终的结果，也无法进行"非理性的创造"。因此，人类恰恰是通过自身的"碳基缺陷"，构筑了不可替代的防线。 "AI的本质是计算，人类的本质是博弈。"陈天桥指出，"选择和承担正是人类最后的防线。"他认为，只有人类作出选择并为结果负责的前提下，AI的运算才 能从"数据处理"真正转化为"商业价值"，他将人类担责的行为称为"确权"。 他还提到马斯克的成功，并称此为"超级个体的崛起"。"马斯克只做两件事：极具野心的选择，指明那个看似不可能的方向；以及赌上一切的担责，承担所 有失败的代价。"陈天桥称，"这就是超级个体的定义。人类现在唯一需要比拼的，就是谁的支点（责任）更稳，谁的施力（选择）更准。" 基于这些观点，陈天桥进一步提出构建"最小可行性责任"的三级路径：建立"人机防火墙"；学会预测成功的概率；拥抱"混乱"，寻找数据缺失的角落，解 决"人心博 ...