October 2025 Summaries

Smit Barmase and Bennet Fenner, two contributors who transitioned from open-source participants to team members at Zed, share their unique journeys and experiences of joining the company. Smit, a former web developer from Bangalore, India, was drawn to the performance-first focus of Zed and contributed significantly to Linux-related features and bug fixes, which led to his hiring. Despite time zone challenges, his work was reviewed asynchronously by the Zed team. Bennet, a Computer Science student from Germany, discovered Zed through a blog post and started contributing after purchasing a MacBook to try the software. He engaged actively with the Zed community and paired with engineers, which facilitated a seamless transition to a formal role at the company. Both found the open-source contribution route rewarding and educational, noting the autonomy and collaborative culture at Zed. Their stories highlight the mutual benefits of hiring through community contributions, as contributors gain real-world experience and relationships, while the company benefits from fresh perspectives and solutions.

The introduction of the Agent Client Protocol (ACP) in collaboration with Google's Gemini CLI team revealed a significant demand for such a protocol, leading to its adoption by various clients and agents, including JetBrains, and prompting requests for OpenAI's Codex integration. Zed now supports Codex through ACP, allowing users to select it from the New Thread menu without any additional charges, as interactions with Codex do not pass through Zed's servers. The ACP is designed for flexibility, accommodating different agent capabilities, but differences in agent implementations, such as Codex's handling of terminal commands, affect user experiences. Codex runs commands in its own process, contrasting with other agents that use pseudoterminal (PTY) mode, which can result in terminal interactivity issues. With the successful integration of multiple agents, Zed is now focusing on collaborating with the community for future ACP developments, having already seen its adoption in editors like Neovim, Emacs, and JetBrains IDEs. The Zed team encourages further community-driven ACP integrations and invites individuals to join their team to continue advancing software development.

Zed is now fully supported on Windows, offering both stable and preview releases, with updates released weekly. The Windows build of Zed integrates directly with the platform, using DirectX 11 and DirectWrite for rendering, and supports Windows Subsystem for Linux (WSL) and SSH remoting for seamless remote editing. Zed's extensions are compatible with Windows and utilize WebAssembly Components for sandboxed access to the file system, eliminating the need for Windows-specific workarounds. All AI features, including edit predictions and ACP-powered agents, are supported on Windows, enhancing the coding experience with integrated tools like Claude Code. Users are encouraged to download and test Zed for Windows, providing feedback on various workflows and features to inform future updates and improvements.

Steve Klabnik discusses the challenges of having productive conversations about AI, emphasizing the importance of understanding other people's interests and being open to learning rather than winning arguments. He highlights the difficulties of such discussions on large social media platforms, suggesting that smaller, more personal settings often yield more fruitful exchanges. Klabnik also notes the importance of recognizing the probabilistic nature of technology and being adaptable in acquiring new skills, as the field is rapidly evolving. He underscores the necessity of being willing to try and fail as a means of skill development, advocating for a mindset that embraces experimentation and continuous learning. The session is part of the Agentic Engineering series, which explores quality software development in an era where code production is less of a constraint.

The blog post explores various underappreciated features and workflows within the Zed editor, highlighting how these tools can enhance productivity and ease of use. With approximately 900 pull requests merged monthly, keeping up with Zed's evolving features can be challenging, but the post shares insights from different team members on useful functionalities. These include keybinding editing to streamline Rust programming, using keyboard shortcuts for repeated test runs, context menu features like "Copy and Trim," and techniques for focusing on prose writing by customizing the workspace layout. The post also discusses advanced search capabilities using regex, the use of multicursors for manipulating text, and the integration of external tools like Lazygit. By sharing these "hidden gems," the article encourages users to explore and share their own tips for maximizing Zed's potential.

Zed's extensive and evolving features make it challenging to keep track of all updates, especially with around 900 pull requests merged monthly. The series highlights lesser-known functionalities and workflows within Zed aimed at enhancing user efficiency. Key features include customizable keybindings to facilitate Rust programming, repeated test runs via keyboard shortcuts, and context menu functions like "Copy and Trim" for cleaner text insertion. Users can benefit from actions like multicursor editing, auto-save on focus change, and symbolic navigation through the outline modal. Zed also offers flexible file navigation options, the ability to open files in new panes, and integrates tools like Lazygit for improved task management. The series encourages users to share their own tips to enrich the community's collective knowledge.

JetBrains has partnered with a team to co-develop the Agent Client Protocol (ACP), aiming to integrate ACP support across its suite of integrated development environments (IDEs) like IntelliJ IDEA, PyCharm, and WebStorm. This collaboration marks a significant step for ACP, which was initially launched by the team alongside Google and Gemini CLI with the goal of establishing an open standard that allows any agent to work with any editor, eliminating the need for custom integrations. With JetBrains' involvement, ACP is positioned to become a standard for agent-editor integration, enhancing developer choice and enabling agents to operate across a variety of platforms without vendor lock-in. The initiative encourages developers to adopt ACP to reach a broader audience, emphasizing the benefits of reduced duplicated work and increased interoperability. This movement towards standardization is gaining traction, with the hope of engaging more agents and editors to adopt ACP, thus advancing the future of software development.

Zed, a code editor, has prioritized improving its Python support this year by enhancing virtual environment detection, managing separate language servers per toolchain, and supporting tools like Pyright and Ruff. The toolchain selector automatically detects and manages Python virtual environments, and Zed is now monorepo-ready, allowing separate toolchains for each sub-project via pyproject.toml tracking. The editor aims to streamline developers' workflow by reducing configuration overhead, a common issue with the Language Server Protocol (LSP) and Python's complex ecosystem. Zed also plans to integrate Ty, a performance-focused server, into its core, while still supporting existing servers like Pyright and Pylsp. Future developments include introducing notebooks in 2026 to further enhance the Python experience, and the team encourages feedback from users to continue refining the tool.

The Agent Client Protocol (ACP) was launched to increase the integration of agents in the Zed editor by providing a universal protocol that any agent can implement, enhancing flexibility and choice for developers. Initial adoption has been promising, with support expanding beyond Zed to other editors like Neovim, Emacs, and marimo notebook, demonstrating its potential across different environments. Key collaborations include Google’s Gemini CLI and Square’s Goose, with ongoing projects for integrating OpenAI's Codex and other agents. The ACP community actively seeks contributions and feedback to improve and expand the protocol, while future developments aim to provide users with greater control over agent configurations.