We Have to Start Over: From Atom to Zed

Original Article ↗ Hacker News Discussion ↗

A new native code editor called Zed, built as a high-performance successor to Atom and alternative to VS Code, is drawing interest for its low latency, GPU-accelerated UI and collaborative “multiplayer” features. Commenters are divided over its Mac-only status, lack of remote development, notebooks and some VS Code-equivalent features, arguing that cross-platform support and ecosystem maturity matter more than shaving milliseconds off UI rendering. Others raise concerns about Zed’s collaboration-related data access terms and the challenges of implementing accessibility in a custom Rust-based UI framework, suggesting these could be decisive for team adoption.

Platform Support & Adoption

  • Major friction: macOS-only release. Many users on Windows/Linux dismiss Zed until ports exist.
  • Linux port is actively developed with strong community contribution; Windows work exists in forks/branches and a Discord channel, but no clear commitment on inclusion timeline.
  • Some users like Mac-only apps for polish and consistency; others highlight the pain of switching between platforms and editors.
  • Remote development (containers/VMs/SSH) is critical for many; lack of this (and devcontainers-style workflows) is a blocker compared with VS Code.

Performance & SIMD

  • Zed is repeatedly praised as noticeably faster than VS Code in startup, file open, and general responsiveness, especially on modern Macs.
  • Some argue VS Code is already “fast enough” for everyday use; performance alone doesn’t justify switching given ecosystem gaps.
  • Discussion on SIMD: some note compilers and standard libraries already use SIMD; others say auto-vectorization often needs deliberate structuring and manual SIMD can still bring big wins.
  • Users compare this to tools like ripgrep showing people they’d normalized unnecessary slowness.

Features, Ecosystem, and “Yet Another Editor”

  • Many feel existing tools (VS Code, JetBrains, Vim/Neovim, Emacs, Sublime, Kate) already cover needs; question what Zed adds beyond performance and multiplayer.
  • Multiplayer/live collaboration is seen as interesting but “edge case” for some; others see it as Zed’s main differentiator.
  • Missing pieces frequently cited: snippets, Jupyter notebooks, robust remote dev, rich extension ecosystem, strong defaults and keyboard navigation, Windows support.
  • Some like specific Zed UX features (e.g., editable global search results with multi-cursor) and its “lightweight editor with IDE power” goal.

Privacy, Licensing, and AI

  • Terms for collaboration: when you share a project, code is stored on Zed’s servers and may be accessed for debugging/improvements.
  • Some see this as standard for hosted services; others call it a red flag, especially for proprietary code, and request explicit opt-ins and/or end-to-end encryption.
  • AI integration is criticized for not allowing a custom OpenAI-compatible host, blocking local LLM use.

Accessibility & Custom UI Framework

  • Concern that Zed’s custom GPU-based UI framework makes accessibility (especially screen readers) hard and currently essentially absent.
  • Multiple commenters stress accessibility must be designed in from the start; retrofitting later is costly and often incomplete.
  • Broader debate on non-native UIs: custom frameworks can deliver speed and control but risk long, under-resourced paths to parity on a11y and UX.