Project ideas from Hacker News discussions.

Zed editor switching graphics lib from blade to wgpu

📝 Discussion Summary (Click to expand)

1. WebGPU/WGPU is a double‑edged sword
- “WebGPU has some surprising performance problems … a well‑made Metal renderer will easily be 100× smaller.” – flohofwoe
- “WGPU is just a layer over the top of the native APIs … it will be slower or at least not much better.” – swiftcoder
- “The WebGPU API gets you to rendering your first triangle quicker … but it lags behind what you can do directly in DX or VK.” – pornel

2. Rust GUI libraries are still immature and fragmented
- “Open source GUI development is perpetually cursed by underestimating the difficulty of the problem.” – api
- “Iced is a passion project by a single developer … it will have to be forked and a community development model built around it.” – tensor
- “GPUI is a low‑level library that Zed built for itself; the community has to build the components.” – the__alchemist

3. Switching Zed’s Linux renderer to WGPU raises memory‑usage and performance questions
- “They felt their native renderer on those platforms was better and less memory intensive.” – nu11ptr
- “Current wgpu seems to have a floor around ~100 mb that isn’t there with other rendering backends.” – nicoburns
- “WGPU is a compatibility layer, adding more code – will it make Zed slower?” – amelius

4. Zed’s value proposition versus established editors (VS Code, nvim, JetBrains)
- “Zed competes mostly against Visual Studio Code. Not against JetBrains.” – qznc
- “Zed is fast, has good project‑wide search, and AI integration, but lacks some IDE‑style refactoring.” – linolevan / kylecazar
- “Zed is a native editor that feels snappy and has AI features, but it still crashes and has limited plugins.” – trcf23 / r4nd0m4ch

These four threads capture the bulk of the discussion: the trade‑offs of adopting WebGPU, the state of Rust GUI tooling, the practical impact of Zed’s renderer change, and how Zed stacks up against other editors.


🚀 Project Ideas

Zed Performance Profiler

Summary

  • A lightweight desktop tool that hooks into Zed’s rendering pipeline to collect real‑time metrics on GPU memory, CPU usage, frame times, and backend switch overhead.
  • Provides actionable recommendations (e.g., switch to native renderer, enable low‑memory mode, adjust texture cache size) to help users reduce RAM footprint and improve responsiveness.

Details

Key Value
Target Audience Zed users, especially those on low‑end hardware or VMs.
Core Feature Live profiling dashboard + post‑run analysis + automated tuning suggestions.
Tech Stack Rust + wgpu + egui for UI, cross‑platform (Linux/Windows/macOS).
Difficulty Medium
Monetization Hobby

Notes

  • Users complained about “Zed 410 MB vs VS Code 580 MB” and “blurry fonts” – a profiler can pinpoint the root causes.
  • “Profiling wgpu vs native” is a hot topic; a ready‑made tool would spark discussion on performance trade‑offs.
  • The tool can be open‑source, encouraging community contributions to the Zed codebase.

Rust GUI Component Library

Summary

  • A fully‑featured, component‑driven Rust GUI framework that supports virtualized lists, accessibility, theming, and high‑performance rendering via wgpu or native backends.
  • Aims to replace the fragmented ecosystem (egui, iced, slint) with a single, mature library.

Details

Key Value
Target Audience Rust developers building desktop or embedded UIs.
Core Feature Rich widget set (tables, trees, forms), virtual scrolling, accessibility APIs, cross‑platform rendering.
Tech Stack Rust, wgpu (or native via winit), parley for text, palette for colors, wasm‑target for web.
Difficulty High
Monetization Revenue‑ready: subscription for enterprise support + open‑source core.

Notes

  • Commenters lament “gpui is low‑level” and “egui lacks virtual lists”; this library directly addresses those gaps.
  • “Rust GUI is in a tough spot” – a unified library would reduce fragmentation and attract contributors.
  • The component model aligns with modern UI patterns, making it attractive for both hobbyists and commercial teams.

Zed Remote Editor Service

Summary

  • A cloud‑hosted Zed instance that runs on inexpensive EC2 or similar, exposing a web‑based editor and SSH/WS backend for file access.
  • Enables developers to edit code from any device without installing Zed locally, while keeping the native performance and AI features.

Details

Key Value
Target Audience Remote developers, CI/CD pipelines, educational platforms.
Core Feature One‑click deployment, secure file system sandbox, WebGPU‑enabled web UI, SSH/WS API for remote editing.
Tech Stack Rust (Zed), Docker, Terraform, AWS EC2, Nginx reverse proxy, TLS.
Difficulty Medium
Monetization Revenue‑ready: pay‑per‑hour or subscription per instance.

Notes

  • “Can this be done on a cheap AWS EC2 instance?” – this service answers that directly.
  • “Zed already has a client/server architecture” – the service leverages that to provide a turnkey remote editor.
  • The model would be useful for code review tools, learning platforms, and distributed teams.

AI Code Completion Plugin for Zed

Summary

  • A plugin that integrates popular LLMs (Claude, GPT‑4, Llama) into Zed, offering context‑aware completions, refactoring suggestions, and AI‑driven navigation.
  • Adds missing IDE features (move function, rename symbol) and improves the AI experience beyond the current “edit prediction”.

Details

Key Value
Target Audience Zed users seeking advanced AI assistance and refactoring tools.
Core Feature Real‑time LLM inference, symbol‑aware refactoring, multi‑model fallback, offline caching.
Tech Stack Rust, Zed plugin API, OpenAI/Anthropic APIs, local LLM inference via llama.cpp or rust‑llm.
Difficulty Medium
Monetization Revenue‑ready: freemium with paid premium models or per‑request billing.

Notes

  • Users noted “Zed lacks refactoring” and “AI integration is weak”; this plugin fills that gap.
  • “Claude Code CLI integration” is a hot topic; the plugin would provide a seamless in‑editor experience.
  • The plugin can be open‑source, encouraging community contributions and model‑agnostic extensions.

Read Later