⌘K
Change language Switch ThemeSign In
Narrow Mode
PingCAP 联合创始人兼 CTO 黄东旭已确认出席 QCon 北京站，分享 The Age of Autonomous Systems 自主系统的时代
I InfoQ 中文 @InfoQ 中文
One Sentence Summary
PingCAP CTO Ed Huang will speak at QCon Beijing on the evolution of software architecture in 'The Age of Autonomous Systems,' exploring the challenges Agents pose to infrastructure and state management as new computing entities.
Summary
This article previews the keynote speech by Ed Huang, Co-founder and CTO of PingCAP, at QCon Beijing. The core argument is that Agents are no longer just functional modules but are becoming a new class of computing entities reshaping the software ecosystem. Huang will focus on the challenges of software acting as an 'active actor' during the evolution from traditional application architecture to 'autonomous system architecture,' including long-term state management, multi-agent collaboration mechanisms, and the infrastructure definitions required to support continuous agent operation, aiming to provide developers with a macro framework for the Agent era.
Main Points
* 1. Agents are evolving from functional modules into a brand-new class of computing entities.This shift is accelerating the evolution of system architecture from traditional application models to 'autonomous system architecture,' irresistibly reshaping the software ecosystem. * 2. Long-term state management is the real challenge facing AI systems.Compared to model capabilities, effectively managing the long-term state of Agents is the key bottleneck in building continuously running intelligent agents and a core dimension of system design. * 3. The shift in software roles requires a redefinition of infrastructure.As software transitions from a passive tool to an 'active actor,' we need to rethink the boundaries of APIs, CLIs, and state-driven design, as well as the foundations that support multi-agent collaboration.
Metadata
AI Score
82
Website mp.weixin.qq.com
Published At Today
Length 1144 words (about 5 min)
Sign in to use highlight and note-taking features for a better reading experience. Sign in now
InfoQ 2026-03-31 13:31 北京
Agent 并非简单的功能模块，而是一种全新的计算实体
从「AI For What」到「Value From AI」，100+ 可落地实践案例打通 AI 实战最后一公里！
4 月 16 日 -4 月 18 日，QCon 全球软件开发大会 将在北京举办。本届大会锚定 Agentic AI 时代的软件工程重塑，聚焦 Agentic AI、多智能体协作、算力优化、技术债治理、多模态和 AI 原生基础设施等前沿话题，邀请来自腾讯、阿里、百度、华为、蚂蚁、小米、网易等企业技术专家，带来百余项真实落地案例，系统性分享前沿洞察与实战干货，以技术共创探索 AI 落地新路径。
PingCAP 联合创始人兼 CTO 黄东旭已确认出席，并将在 Keynote 主题演讲 中发表题为《The Age of Autonomous Systems 自主系统的时代》的主题分享。2026 年初，随着 OpenClaw 等技术的出现，业界已深刻意识到：Agent 并非简单的功能模块，而是一种全新的计算实体。它正以不可阻挡之势重塑软件生态，推动系统架构从传统应用架构，加速向 “自主系统架构” 演进。 对此，在本次演讲中，黄东旭将围绕几个关键问题展开：
* 当软件成为“主动行动者”，系统架构如何演化？
* 为什么 AI 的真正挑战不是模型能力，而是长期状态管理？
* 从 API 到 CLI 到状态驱动，边界在哪里？
* 多 Agent 协作将如何重塑软件系统？如何有效的实现协同？
* 什么样的基础设施能够支撑持续运行的智能体？
黄东旭，PingCAP 联合创始人兼 CTO。开源分布式数据库 TiDB 项目发起人，专注于分布式系统、数据库内核与云原生架构领域十余年。TiDB 作为全球领先的开源 HTAP 数据库，已在金融、互联网、电信等行业数千家企业落地，支撑海量数据的实时处理与分析。近年来深入研究 AI 基础设施与 Agent 系统架构，探索大模型时代的数据存储与计算范式变革。长期活跃于开源社区，曾在多个国际技术会议发表演讲，致力于推动数据库技术的创新与普及。
通过本次演讲，听众将获得：
* 对 Agent 时代的宏观理解框架
* 对未来软件架构演化方向的判断
* 对长期状态系统设计的思考维度
* 对“Agent Infra”真正含义的重新定义
除此之外，本次大会还策划了 Agentic Engineering、多模态理解与生成的突破、记忆觉醒：智能体记忆系统的范式重塑与产业落地、具身智能与物理世界交互、Agent Infra 架构设计、AI 重塑数据生产与消费、AI 原生基础设施、AI 驱动的技术债治理、小模型与领域适配模型、大模型算力优化、Agent 可观测性与评估工程、AI for SRE 等 20 多个专题论坛，届时将有来自不同行业、不同领域、不同企业的 100+ 资深专家在 QCon 北京站现场带来前沿技术洞察和一线实践经验。
会议推荐
QCon 全球软件开发大会·2026 北京站 将于4 月 16 日-18 日正式举办。本届大会以“Agentic AI 时代的软件工程重塑”为主题，聚焦100+重磅议题，汇聚来自阿里、腾讯、字节跳动、小米、百度等一线科技企业与创新团队的技术专家，围绕 AI 工程化、系统架构与研发模式演进展开深入探讨。更多详情可扫码或联系票务经理 18514549229 进行咨询。
I InfoQ 中文 @InfoQ 中文
One Sentence Summary
PingCAP CTO Ed Huang will speak at QCon Beijing on the evolution of software architecture in 'The Age of Autonomous Systems,' exploring the challenges Agents pose to infrastructure and state management as new computing entities.
Summary
This article previews the keynote speech by Ed Huang, Co-founder and CTO of PingCAP, at QCon Beijing. The core argument is that Agents are no longer just functional modules but are becoming a new class of computing entities reshaping the software ecosystem. Huang will focus on the challenges of software acting as an 'active actor' during the evolution from traditional application architecture to 'autonomous system architecture,' including long-term state management, multi-agent collaboration mechanisms, and the infrastructure definitions required to support continuous agent operation, aiming to provide developers with a macro framework for the Agent era.
Main Points
* 1. Agents are evolving from functional modules into a brand-new class of computing entities.
This shift is accelerating the evolution of system architecture from traditional application models to 'autonomous system architecture,' irresistibly reshaping the software ecosystem.
* 2. Long-term state management is the real challenge facing AI systems.
Compared to model capabilities, effectively managing the long-term state of Agents is the key bottleneck in building continuously running intelligent agents and a core dimension of system design.
* 3. The shift in software roles requires a redefinition of infrastructure.
As software transitions from a passive tool to an 'active actor,' we need to rethink the boundaries of APIs, CLIs, and state-driven design, as well as the foundations that support multi-agent collaboration.
Key Quotes
* Agents are not just simple functional modules, but a brand-new class of computing entities. * Why is the real challenge of AI not model capability, but long-term state management? * System architecture is accelerating its shift from traditional application architecture to 'autonomous system architecture' with unstoppable momentum.
AI Score
82
Website mp.weixin.qq.com
Published At Today
Length 1144 words (about 5 min)
Tags
Agentic AI
Autonomous Systems
Software Architecture
State Management
AI Infrastructure
Related Articles
* Jensen Huang's GTC 2026 Keynote Transcript: All SaaS Companies Will Disappear; Lowest Token Cost Globally; "OpenClaw" Made History; Feynman Architecture is on the Way * Practices and Reflections on Vibe Coding in Code Generation and Collaboration * Why Software Becomes Increasingly Complex: A Deep Dive into Software 'Complexity' * OpenAI Frontline Development Observations: Those Who Can Manage 10-20 Agents Simultaneously and Run Hour-Long Tasks Are Leaving Other Engineers Far Behind * “AI on the Front Lines: How Developers are Reshaping the Software Development Process” | Roundtable Discussion * From Context to Long-Term Memory: Architectural Design and Practice of LLM Memory Engineering\" architecture.") * Harness Engineering: Leveraging Codex in an Agent-First World * 1,500 PRs, 0 Human Coders: Building a Million-Line Internal Product Driven by Codex * Major Update! Jensen Huang's First Keynote of the Year: The World is Undergoing Two Platform Revolutions! New Rubin Computing Architecture Debuts! First Autonomous Driving AI Open-Sourced, AI-Native Application Framework Released! Grok 5 Parameter Scale Revealed * You Might Not Need Workflow Anymore, Skills Are Sufficient for Most Scenarios — A Five-Step Framework to Transform Workflow into Evolvable Skills HomeArticlesPodcastsVideosTweets