视频 · 前往原文观看几年前我们刚开始构建 Cursor 时，大多数代码还是一个按键一个按键敲出来的。Tab 自动补全改变了这一点，开启了 AI 辅助编程的第一个时代。
随后智能体出现了，开发者转而通过同步的提示词-响应循环来指挥智能体。那是第二个时代。现在第三个时代正在到来。这个时代的标志是，智能体能够在更长的时间尺度上、以更少的人类指导，独立处理更大的任务。
因此，Cursor 的核心定位已不再是编写代码。它致力于帮助开发者构建创造软件的"工厂"。这个工厂由一群智能体组成，开发者像与队友协作一样与它们互动：提供初始方向、为它们配备独立工作的工具，并审查它们的工作成果。
我们 Cursor 团队中的许多人已经在以这种方式工作。我们合并的 PR 中，超过三分之一现在是由运行在云端自有计算机上的智能体创建的。我们认为，一年后，绝大多数开发工作将由这类智能体完成。
从 Tab 到智能体
Tab 擅长识别那些低熵、重复性的工作可以在哪里实现自动化。在将近两年的时间里，它带来了显著的效率提升。
然后模型进步了。智能体能够承载更多上下文、使用更多工具，并执行更长的动作序列。开发者的习惯开始转变——夏天时还比较缓慢，但在过去几个月里，随着 Opus 4.6、Codex 5.3 和 Composer 1.5 的发布，转变速度急剧加快。
这种转变已经如此彻底，以至于如今大多数 Cursor 用户几乎不再触碰 Tab 键。2025 年 3 月，我们的 Tab 用户数量大约是智能体用户数量的 2.5 倍。现在，这个比例完全反转了：智能体用户数量是 Tab 用户数量的 2 倍。
但这一转变已经在让位于更重大的变革。Tab 时代持续了将近两年。第二个时代——即大部分工作由同步智能体完成的时代——可能连一年都持续不了。
云端智能体与工件
与 Tab 相比，同步智能体在更上层的技术栈中工作。它们处理需要上下文和判断力的任务，但仍让开发者在每一步都参与其中。但这种实时交互的形式，加上同步智能体在本地机器上竞争资源的事实，意味着同时只与少数几个智能体协作才是可行的。
云端智能体消除了这两个限制。每个智能体都在自己的虚拟机上运行，允许开发者移交任务后转而处理其他事情。智能体花费数小时逐步完成任务，反复迭代和测试，直到对输出结果有信心，然后返回可供快速审查的内容：日志、视频记录和实时预览，而非代码差异对比。
这使得并行运行智能体变得可行，因为产物和预览提供了足够的上下文来评估输出，而无需从头重建每个会话。人类的角色从指导每一行代码转变为定义问题和设定审查标准。
这一转变正在 Cursor 内部发生
在 Cursor 内部，我们合并的 PR 中有百分之三十五现在是由在云端虚拟机中自主运行的智能体创建的。我们看到采用这种新工作方式的开发者具有三个特征：
- 智能体编写了他们几乎百分之百的代码。
- 他们把时间花在分解问题、审查产物和提供反馈上。
- 他们同时启动多个智能体，而不是手把手地指导一个智能体直到完成。
在软件开发中，这种方法要成为标准，还有很多工作要做。在工业规模上，单个开发者可以绕过的不可靠测试或损坏环境，会变成中断每个智能体运行的故障。更广泛地说，我们仍需确保智能体能够尽可能高效地运行，拥有它们所需的全部工具和上下文。
我们认为昨天发布的 Cursor 云端智能体是朝着这个方向迈出的初步但重要的一步。
想法
提示词设计
Arvid Lunnemark
视频 · 前往原文观看When we started building Cursor a few years ago, most code was written one keystroke at a time. Tab autocomplete changed that and opened the first era of AI-assisted coding.
Then agents arrived, and developers shifted to directing agents through synchronous prompt-and-response loops. That was the second era. Now a third era is arriving. It is defined by agents that can tackle larger tasks independently, over longer timescales, with less human direction.
As a result, Cursor is no longer primarily about writing code. It is about helping developers build the factory that creates their software. This factory is made up of fleets of agents that they interact with as teammates: providing initial direction, equipping them with the tools to work independently, and reviewing their work.
Many of us at Cursor are already working this way. More than one-third of the PRs we merge are now created by agents that run on their own computers in the cloud. A year from now, we think the vast majority of development work will be done by these kinds of agents.
From Tab to agents
Tab excelled at identifying where low-entropy, repetitive work could be automated. For nearly two years, it produced significant leverage.
Then the models improved. Agents could hold more context, use more tools, and execute longer sequences of actions. Developer habits began to shift, slowly through the summer, then rapidly over the last few months with the releases of Opus 4.6, Codex 5.3, and Composer 1.5.
The transformation has been so complete that today, most Cursor users never touch the tab key. In March 2025, we had roughly 2.5x as many Tab users as agent users. Now, that is flipped: we now have 2x as many agent users as Tab users.
But already this shift is giving way to something bigger. The Tab era lasted nearly two years. The second era, in which most work is done with synchronous agents, may not last one.
Cloud agents and artifacts
Compared to Tab, synchronous agents work further up the stack. They handle tasks that require context and judgment, but still keep the developer in the loop at every step. But this form of real-time interaction, combined with the fact that synchronous agents compete for resources on the local machine, means it is only practical to work with a few at a time.
Cloud agents remove both constraints. Each runs on its own virtual machine, allowing a developer to hand off a task and move on to something else. The agent works through it over hours, iterating and testing until it is confident in the output, and returns with something quickly reviewable: logs, video recordings, and live previews rather than diffs.
This makes running agents in parallel practical, because artifacts and previews give you enough context to evaluate output without reconstructing each session from scratch. The human role shifts from guiding each line of code to defining the problem and setting review criteria.
The shift is underway inside Cursor
Thirty-five percent of the PRs we merge internally at Cursor are now created by agents operating autonomously in cloud VMs. We see the developers adopting this new way of working as characterized by three traits:
- Agents write almost 100% of their code.
- They spend their time breaking down problems, reviewing artifacts, and giving feedback.
- They spin up multiple agents simultaneously instead of handholding one to completion.
There is a lot of work left before this approach becomes standard in software development. At industrial scale, a flaky test or broken environment that a single developer can work around turns into a failure that interrupts every agent run. More broadly, we still need to make sure agents can operate as effectively as possible, with full access to tools and context they need.
We think yesterday's launch of Cursor cloud agents is an initial but important step in that direction.
Ideas
Prompt design
Arvid Lunnemark