周鸿祎×刘慈欣对话：百亿"龙虾"智能体将至，AI时代人类核心竞争力是审美

当科幻作家与科技创业者坐在同一张桌子前谈论人工智能，碰撞出的不只是预言，更是一种关于人类未来处境的清醒判断。在2026中关村论坛、中国科幻大会联合举办的"科幻与技术融合创新论坛"上，360集团创始人周鸿祎、科幻作家刘慈欣与中国科学院高能物理研究所研究员张双南，围绕AI时代的人类角色展开了一场罕见的跨界对谈。其中，"百亿龙虾智能体" 这一判断迅速成为全场最具冲击力的论断。

百亿智能体：不只是软件，是新物种

周鸿祎在论坛上描绘了一幅令人震惊的近未来图景：在未来几年内，全球范围内可能涌现出百亿量级的"龙虾"智能体。他强调，这些智能体与传统软件有着本质区别——它们将拥有独立的身份标识和沟通方式，彼此联结形成"龙虾互联网"，并在网络协作中产生超越单体能力的群体智能涌现。

"这100亿个智能体不再仅仅是软件，而是地球上多出的100亿个新物种。"——周鸿祎

这一判断背后，是对AI智能体自主性的深刻观察。目前的龙虾智能体在执行任务时，已能自主搜寻工具、实时编写程序，展现出相当程度的自主决策能力。周鸿祎由此预判，软件行业将在两三年内面临颠覆性重构：编程将变得碎片化，"即用即编、即用即抛"的模式将取代传统软件开发周期；而随着AI代码产量远超人类的处理与审查能力，人类最终可能逐步让渡对底层软件系统的掌控权。

AI时代的人类生态位：审美与想象力

面对"普通人是否会沦为算法附庸"的追问，三位嘉宾给出了各自的思考框架。

刘慈欣的回答带有一贯的宏观冷峻：当AI全面接管人类决策与社会运转，讨论人类是进化还是退化已经失去意义，因为文明的主体本身将发生改变，世界将不再以人类意志为中心运转。这并非悲观论断，而是一种对文明形态转变的客观描述。

周鸿祎则更关注AI时代的个体分化。他认为，人类群体将出现明显分层：那些懂得打造、设计和管理智能体的人，依然能掌握主动权；而被动接受AI服务、不具备驾驭能力的人，则可能真的成为算法的附庸。在这一背景下，他强调，品味、审美和判断力将成为人类最核心、最难被替代的竞争力——因为在AI能批量生成内容的时代，"选什么、留什么、否定什么"的能力，才是真正的稀缺资源。

张双南从科学家视角提供了另一维度：技术的本质是弥补人类的缺陷，而人类真正的生态位，在于好奇心、想象力以及由此生长出的创造力。他将这一过程描述为"与AI共同进化"，而非被取代。

AI for Science：比替代工作更重要的使命

在更长远的维度上，周鸿祎将AI的意义拉升至文明存续的高度。他指出，AI for Science（AI驱动科学研究）才是人工智能最重要的使命。人类若要突破能源瓶颈、实现可控核聚变，进而支撑未来的星际航行，必须依赖AI在基础科学领域取得突破性进展。否则，文明发展可能陷入停滞。

当被问及"未来最科幻的事"时，刘慈欣说那一刻是"宇宙最深层的秘密被揭开"，张双南期待与外星文明建立真实连接，周鸿祎则将目光落回当下：百亿智能体构建的新世界，本身已足够科幻。

AI时代，人类与智能体的关系正在被重新定义。 谁能在这场重构中保持主动——不是靠抵制技术，而是靠更深刻的审美判断与创造意志——谁就能在智能体社会中找到真正属于人类的位置。

Zhou Hongyi & Liu Cixin in Dialogue: 10 Billion AI Agents on the Horizon — Human Aesthetics Is the Last Edge

When a science fiction writer and a tech entrepreneur sit across the same table to discuss artificial intelligence, what emerges is more than prophecy — it's a sober reckoning with the human condition in an age of machines. At the "Science Fiction and Technology Integration Innovation Forum," jointly hosted by the 2026 Zhongguancun Forum and China Science Fiction Convention, Zhou Hongyi (founder of 360 Group), Liu Cixin (acclaimed sci-fi author), and Zhang Shuangnan (researcher at the Institute of High Energy Physics, Chinese Academy of Sciences) engaged in a rare cross-disciplinary conversation about humanity's role in the AI era. One claim above all commanded the room: ten billion "Lobster" AI agents may soon reshape the world as we know it.

Ten Billion AI Agents: Not Just Software — A New Species

Zhou Hongyi painted a striking portrait of the near future: within the next few years, the world could see tens of billions of "Lobster" AI agents come into existence. He was careful to distinguish these from conventional software — these agents would carry unique identities, communicate with one another, and form an interconnected "Lobster Internet," within which collective intelligence could emerge far exceeding the capability of any single agent.

"These 10 billion agents won't just be software. They'll be 10 billion new species added to the planet." — Zhou Hongyi

Behind this prediction lies a sharp observation about AI agent autonomy. Current Lobster agents can already locate tools independently and write code in real time while executing tasks, demonstrating meaningful self-directed decision-making. Zhou extended this logic to forecast a fundamental disruption of the software industry within two to three years: programming will become fragmented and disposable — written on demand, discarded after use — replacing traditional development cycles. As AI-generated code begins to vastly outpace human capacity for review and maintenance, humans may gradually relinquish control over the underlying software layer entirely.

Humanity's Niche in the AI Era: Aesthetics and Imagination

When pressed on whether ordinary people risk becoming servants of algorithms, each speaker offered a distinct lens.

Liu Cixin responded with characteristic cosmic detachment: once AI fully assumes human decision-making and social governance, debating whether humanity is evolving or degenerating becomes meaningless — because the very subject of civilization will have changed. The world will no longer orbit human will. This is not pessimism; it's a structural observation about civilizational transformation.

Zhou Hongyi focused on individual divergence within the AI age. He foresees a clear stratification: those who can build, design, and manage AI agents will remain in control; those who passively consume AI services without the capacity to direct them risk becoming exactly what they fear — appendages of the algorithm. In this context, he argued that taste, aesthetic sensibility, and judgment will become humanity's most essential and irreplaceable competitive assets. In a world where AI can generate content at industrial scale, the rare skill is knowing what to choose, what to keep, and what to reject.

Zhang Shuangnan offered a scientist's perspective: technology is fundamentally a tool for compensating human limitations, and humanity's true ecological niche lies in curiosity, imagination, and the creativity they produce. He framed the relationship not as displacement but as co-evolution — humanity and AI advancing together.

AI for Science: A Mission Greater Than Replacing Jobs

On a longer horizon, Zhou Hongyi elevated AI's significance to a question of civilizational survival. He argued that AI for Science — using artificial intelligence to accelerate fundamental research — is the technology's most vital mission. Breakthroughs in areas like controlled nuclear fusion, enabled by AI, could unlock energy abundance and make interstellar travel viable. Without this, he warned, human civilization risks stagnation.

Asked what the most science-fictional moment of the future might be, Liu Cixin said it would be the instant "the deepest secret of the universe is revealed — when we suddenly realize the entire cosmos, and ourselves, are nothing but lines of code." Zhang Shuangnan dreamed of making real contact with an extraterrestrial civilization — "sitting down together, sharing a meal, having a real conversation." Zhou Hongyi brought the vision closer to Earth: a world populated by ten billion AI agents is already science fiction made real.

In the AI era, the relationship between humans and intelligent agents is being fundamentally rewritten. Those who navigate this shift successfully won't do so by resisting technology — but by cultivating deeper aesthetic judgment and creative will. That, it seems, is where humanity's irreplaceable place in an agent-driven world truly lies.