transcript
show notes
Chinese artificial intelligence startup DeepSeek has unveiled a new architecture aimed at making AI agents more autonomous, marking the latest push by China's model developers to move beyond chatbots toward systems capable of independently carrying out complex tasks.
中国人工智能初创企业深度求索（DeepSeek）推出一套全新架构，旨在提升智能体的自主能力。这标志着国内大模型研发团队的最新探索方向：不再局限于对话机器人，转而打造能够独立完成复杂任务的系统。
DeepSeek's Harness team said in an article published on Saturday that its experimental "super-agent" architecture is designed to allow AI to continuously monitor information, manage multiple tasks in parallel and take action with less human supervision.
深度求索旗下 Harness 团队在周六发布的文章中介绍，这套实验性 “超级智能体” 架构，可以让人工智能持续监测信息、并行处理多项任务，在更少人工干预下自主执行操作。
The development points to a broader shift in China's fast-moving AI industry, where competition is increasingly moving beyond model size and benchmark scores toward agents that can plan, execute and complete real-world work.
这项成果折射出国内高速发展的人工智能行业正在发生重要转向：行业竞争不再只比拼模型参数规模与基准测试分数，重心转向能够规划、执行并落地真实业务的智能体。
As AI models become better at handling individual tasks, DeepSeek said the bottleneck is increasingly human attention: users still have to switch between different conversations, keep track of their progress and provide instructions.
深度求索表示，随着大模型处理单项任务的能力持续提升，瓶颈逐渐转移到人力投入上。用户需要在多个对话窗口之间切换、跟进任务进度、持续下发指令。
The super-agent approach is intended to move much of that coordination into the AI system itself, allowing people to focus more on higher-level decisions and reviewing results rather than overseeing each step of a task.
超级智能体方案，就是把大量协调工作交由 AI 系统自主完成。用户可以更多聚焦顶层决策与结果核验，不必盯着任务的每一个执行环节。
DeepSeek stressed that the architecture remains an engineering exploration rather than a feature available in its existing products.
深度求索强调，该架构目前仍属于工程探索项目，还没有作为功能上线到现有产品中。
DeepSeek said an internal version of the system is already operating continuously. When employees flag a software problem in a group chat, for example, the AI can identify the potential defect even without being directly addressed, attempt to reproduce the problem in the background and search for a fix.
深度求索称，这套系统的内部版本已经在持续运行。举例来说，员工在群聊里反馈软件故障时，即便没有专门 @它，AI 也能识别潜在缺陷，在后台复现问题，并寻找解决方案。
The push reflects a growing focus across the AI industry on so-called agentic AI — systems designed not only to generate text or answer questions, but also to make plans, use tools and carry out multi-step tasks.
这一研发方向，也体现整个 AI 行业对 “智能体 AI” 日益重视。这类系统不只是生成文本、回答问题，还可以制定方案、调用工具、完成多步骤任务。
Chinese developers have accelerated model releases in recent weeks. The flurry of releases suggests China's AI race is entering a new phase.
近几周，国内厂商加快了大模型发布节奏。密集上新，意味着中国人工智能竞赛进入新阶段。
Rather than simply building models that are bigger or score higher on tests, developers are increasingly competing over whether AI can reliably turn intelligence into action.
研发方的竞争重点不再是单纯做大模型、刷高测试分数，而是比拼 AI 能不能稳定地把推理能力转化为实际行动。
Moonshot AI released last month Kimi K3.1 Pro, the first model in its new K3.1 series.
月之暗面（Moonshot AI）上月发布 Kimi K3.1 Pro，这是其全新 K3.1 系列的首款模型。
The company said the model, its most capable in the roughly 3-trillion-parameter class, offers substantial improvements in coding, 3D creation, financial research and other long-horizon agent tasks, while supporting a context window of 1 million tokens.
该企业介绍，这款模型是其约 3 万亿参数梯队里能力最强的版本，在代码编写、3D 生成、金融调研以及长周期智能体任务上有大幅提升，支持 100 万上下文窗口。
MiniMax launched its M3.1 Flash model mid-October, saying it had strengthened performance in agentic tasks, coding and computer use while delivering online inference speeds nearly three times faster than its predecessor and cutting the overall cost of agent tasks by 70 percent.
稀宇科技（MiniMax）在 10 月中旬推出 M3.1 Flash 模型。官方称，该模型强化了智能体任务、代码、电脑操作能力，在线推理速度达到前代产品近三倍，智能体任务总成本降低 70%。
DeepSeek itself updated its V4.1 Flash model on Oct 17. DeepSeek said the update improved performance while allowing tasks to be completed faster and with shorter outputs.
深度求索也在 10 月 17 日更新了 V4.1 Flash 模型。深度求索表示，本次更新优化了模型能力，任务完成速度更快，输出内容也更加精简。
Wei Kai, head of the AI research institute at the China Academy of Information and Communications Technology, said in a separate interview that agents represent an engineering enhancement built on top of large language models that can significantly extend their capabilities.
中国信息通信研究院人工智能研究所所长魏凯在另一场采访中表示，智能体是在大语言模型基础上做的工程化增强，可以大幅拓展模型能力边界。
But challenges remain. "Agents still need improvement in reliability, contextual memory and long-horizon tasks, and there is still some distance to go before large-scale adoption," Wei said.
但挑战依旧存在。魏凯说：“智能体在稳定性、上下文记忆、长周期任务方面仍有待提升，距离大规模落地应用还有一段路要走。”
autonomous /ɔːˈtɒnəməs/adj. 自主的
architecture /ˈɑːkɪtektʃə(r)/n. 架构；体系
benchmark /ˈbentʃmɑːk/n. 基准；评测标准
inference /ˈɪnfərəns/n. 推理；（模型）推理运算