惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

Microsoft Azure Blog
Microsoft Azure Blog
有赞技术团队
有赞技术团队
Threat Intelligence Blog | Flashpoint
Threat Intelligence Blog | Flashpoint
cs.CV updates on arXiv.org
cs.CV updates on arXiv.org
F
Fox-IT International blog
Recorded Future
Recorded Future
T
ThreatConnect
T
The Exploit Database - CXSecurity.com
SecWiki News
SecWiki News
奇客Solidot–传递最新科技情报
奇客Solidot–传递最新科技情报
人人都是产品经理
人人都是产品经理
T
Tenable Blog
L
LINUX DO - 最新话题
博客园_首页
Hugging Face - Blog
Hugging Face - Blog
罗磊的独立博客
博客园 - 司徒正美
The Hacker News
The Hacker News
博客园 - 聂微东
CTFtime.org: upcoming CTF events
CTFtime.org: upcoming CTF events
Scott Helme
Scott Helme
博客园 - 【当耐特】
O
OpenAI News
Schneier on Security
Schneier on Security
Latest news
Latest news
S
Security @ Cisco Blogs
S
Secure Thoughts
F
Full Disclosure
L
Lohrmann on Cybersecurity
S
SegmentFault 最新的问题
T
Tor Project blog
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
量子位
小众软件
小众软件
T
Threat Research - Cisco Blogs
Simon Willison's Weblog
Simon Willison's Weblog
IT之家
IT之家
大猫的无限游戏
大猫的无限游戏
N
News and Events Feed by Topic
E
Exploit-DB.com RSS Feed
J
Java Code Geeks
Last Week in AI
Last Week in AI
酷 壳 – CoolShell
酷 壳 – CoolShell
Application and Cybersecurity Blog
Application and Cybersecurity Blog
S
Schneier on Security
Cisco Talos Blog
Cisco Talos Blog
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More
P
Proofpoint News Feed
Recent Commits to openclaw:main
Recent Commits to openclaw:main
雷峰网
雷峰网

DEV Community

Karpathy's LLM Wiki? No Code with Claude or Github Copilot! Why Platform Governance and Transparency Matter for Developers and Freelancers Using an LLM to automate a task that used to take hours by hand CyberArena – Interactive Cyber Security Simulation & Threat Analysis Platform Mathematical Functions in CSS: clamp, min, max and How They Simplify Responsiveness Polyglot Persistence in Microservices: Let the Domain Choose the Database 190 Countries, Zero API Calls: Shipping Static Data in a Chrome Extension Your AI Writes Code Fast. Here’s How to Check It Before Shipping qwen2.5-coder is too slow for Claude Code on a Mac. Here's the fix. Building Automated Text-to-Video Pipelines with AI Can Gemini Become an Offline AI Tutor? Lessons from Building Educational AI OPRIX : From a simple messaging web app to a well structured and enhanced UI messaging web app Why React + TypeScript Nullability Slowly Becomes Exhausting Why AI Agents Need a Project Layer - Part 1 Stop Hand-Editing MCP Configs: A Zero-Dependency Go CLI What I Learned Working With Microsoft, SQUAD(GTCO), and Different Tech Communities 🧠 Hermes Agent Assistant — A Modular AI Agent System with Planner, Executor & Memory Spring Boot Auto-Configuration Source Code: Nail This Interview Question The Ultimate Guide to Free AI API Keys: 6 Platforms You Need to Know Why 91% of AI Agents Fail in Production (And What the 9% Do Differently) TryHackMe | Battery | WALKTHROUGH Stop Guessing Your Regex — Test It Live in the Browser I Built FreelancEye, an Open-Source Mobile PWA for Finding Clients Beyond the Hype: My Production Playbook for Docker Swarm Top AI App Builder Platforms with Integrated Backend, Hosting & Database ECS vs EKS in 2026: An Honest Comparison from Someone Who Has Run Both in Production Hardening Your Node.js App Against Supply Chain & Remote Code Execution Attacks linux commands A Practical GEO Case: How an AI System Started Recommending Our Blog Your AI Agent Works 24/7 and Earns $0. I Built the Fix. Your AI Trading Agent Will Lose All Your Money — Here's How To Stop It Google I/O 2026: What Happens When Everything Connects? Why AI writes software but doesn’t build a good product Beyond the Hype: How Google I/O 2026 Secretly Democratized Production-Ready AI Agents with Managed Sandboxes. The Killer Assumption Test: How to Spot Doomed Product Decisions Before You Ship Stop Describing Your Bugs — Just Screenshot Them # I Built an AI Website Builder and Here's What Actually Happened Cooking an AI Campaign in 5 Minutes with Google Cloud AI APIs Your PM Retrospectives Are Lying to You How I Built a Free, Self-Hosted Pipeline That Auto-Generates Faceless YouTube Shorts TypeScript 54 to 58: The Features That Actually Matter in 2026 How to Tailor Your CV to Any Job Posting in 2026 The 7-day SaaS MVP loop: ship fast, then validate with people who actually show up 95. Fine-Tuning LLMs: Make a General Model Do Your Specific Job What Is a Frontend Developer Roadmap and Why You Need One Google shipped three Gemini "Flash" models. Picking the wrong one could 6 your AI bill Building an MCP server so Claude can query my SaaS analytics directly Google I/O 2026 and the Rise of the AI Ecosystem Your Docker Builds Are Slow Because You're Doing It Wrong (And I Built a Tool to Prove It) How do you verify GitHub contributions without trusting self-reported skills? CV vs Resume: What's the Difference and Which Do You Need? student Devs: Build AI Agents & Compete for $55K in Prizes 🚀 How to Write a Cover Letter That Actually Gets You Interviews Battle-Tested: What Getting Hacked Taught Me About Web & Cyber Security Unda folders za kuandika code >> mkdir src >> cd src >> mkdir controllers database routes services utils >> cd .. Directory: C:\Users\mwaki\microfinance-system Mode LastWriteTime Length Name Code Coverage .NET AI slop debt" is technical debt on fast forward. Nobody's ready. Multi-Head Latent Attention (MLA) Memoria - A Local AI Reading Companion Powered by Gemma 4 Stop Trusting Your Accuracy Score: A Practical Guide to Evaluating Logistic Regression Models Serious Question: Is the Developer Job Actually in Risk Due to AI? published: true tags: #discuss #career #ai #help rav2d: We ported an AV2 video decoder from C to Rust — here's why Your New Domain's First Week of GA4 Is a Lie: 4 Days of Raw Data from a Launch Gemma Guide - Real-Time Spatial Awareness for Blind Users From YAML to AI Agents: Building Smarter DevOps Pipelines with MCP A Field Guide to Human–AI Relations (For the Newly Bewildered Mortal) The AI Agent That Learns While It Works — A Complete Guide to Hermes Agent Inviting collaborators to work on ArchScope ArchScope is an interactive web-based tool that lets you design, visualize, and test system architectures with real-time performance simulations. Github - ArchScope is an interactive web-based tool that lets you Gemma 4: Google's Open-Weight AI Is a Game Changer for Developers Confessions of a Git Beginner: Why the Terminal Stopped Scaring Me Docker 容器化实战:从零到生产部署 🚀 I Built a Full Stack Miro Clone with Real-Time Collaboration using Next.js Building an African Economic Data Pipeline with Python, DuckDB & World Bank API llms.txt vs robots.txt vs ai.txt: The Developer's Cheat Sheet Intigriti Challenge 0526 Writeup Business Logic Flaws: How Attackers Skip Steps in Your App to Get What They Should Never Have Why Vibe Coders Need Boilerplates to Save Time, Tokens, and Build More Secure SaaS Projects Idle Cloud Cost Is the New Egress Cost Quark's Outlines: Python Traceback Objects Ghost in the Stack (Part 1): Why uninitialized variables remember old data Building a High-Performance Local Chess Assistant Extension with WebAssembly Stockfish and Manifest V3 Breaking the Trade-off Between Self-Custody and Intelligent Automation on the Stellar Network I Open-Sourced a Practical Fullstack Interview Preparation Repository (React + Node + System Design) 🚀 How I Started Coding as a Student (Beginner-Friendly Guide) WordPress vs. Ghost: Why Automated Bot Attacks Are Making us think much I tested 4 AI agent-governance tools against an open spec - here's the matrix zkML Inference Proof: What the Receipt Proves, and What the Model Still Does Not I Scored 1000/1000 on AWS Certified AI Practitioner (AIF-C01) Here's Every Resource I Used Go - Struct and Interface Handling JSON Requests in Go Storing Kamal secrets in AWS Secrets Manager and deploying to a cheap Hetzner VPS How I Caught and Fixed an N+1 Query in My Django REST API I got tired of paying $10/month to remove image backgrounds – so I built it for free How to Start Coding as a Student: A Complete Beginner’s Guide 🚀 Storing Kamal secrets in AWS Secrets Manager and deploying to a cheap Hetzner VPS What Are Buffers? Build AI Agents with Hot Dev The Client Onboarding Checklist That Prevents 90% of Project Problems Scalable Treasure Hunts Are a Myth, But We Almost Made One Gemini 3.5 Flash Has a 1M Token Context Window. Here's What You Can Actually Build With It.
Why Zed Is Replacing VS Code in My AI-Augmented Workflow
v. Splicer · 2026-05-23 · via DEV Community

VS Code won the editor wars by becoming the Walmart of IDEs — everything available, nothing optimized. That worked fine until AI coding assistants started making the underlying editor speed matter again. When your workflow involves spinning up Claude Code sessions, running agents in the terminal, and context-switching fast between files and inline edits, the latency VS Code buries in its Electron shell adds up in ways you start noticing.

I switched to Zed about four months ago. Not as an experiment. It stuck.


The Architecture Problem VS Code Has Always Had

VS Code runs on Electron. That means Chromium, Node.js, and a JavaScript UI layer between you and your code. The performance is acceptable until you're doing multiple things at once — a language server running, a terminal process active, an AI suggestion loading, a second file open in a split view. At that point you're paying an overhead tax on every action.

Zed is written in Rust. It uses GPUI, a custom GPU-accelerated UI framework the Zed team built themselves. The editor renders to the GPU directly. That's not a marketing claim — you can feel it in scroll performance, file switching, and cursor response. On the same machine, the same files, the same tasks, Zed is noticeably faster.

This matters more now than it did three years ago. AI-assisted coding means the editor is doing more per session: streaming inline completions, rendering diffs, managing multi-file context. Every millisecond of unnecessary latency compounds.


Multiplayer and Collaboration That Doesn't Feel Bolted On

Zed was built with multiplayer from the ground up. Two people can edit the same file simultaneously with real-time cursor visibility, which is what you'd expect. What's less obvious is that the architecture that enables this also makes the whole editor snappier for single-user work — the concurrency model is built into the core rather than being an extension layer.

For solo work, the multiplayer infrastructure mostly stays out of the way. For pair sessions or async collaboration on a security research writeup or a shared script, it's actually useful without requiring a plugin, an account handshake, or a service tier.

LiveShare in VS Code always felt like a plugin duct-taped to something that wasn't designed for it. In Zed it's structural.


The AI Layer: Where the Switch Actually Made Sense

This is the part that moved me. Zed has native AI integration — not through an extension, but built into the editor's architecture. The AI panel, inline completions, and the assistant pane are first-class surfaces.

You can run Claude as your AI backend in Zed directly. Configure it in your settings.json:

{
  "assistant": {
    "default_model": {
      "provider": "anthropic",
      "model": "claude-sonnet-4-20250514"
    },
    "version": "2"
  }
}

Enter fullscreen mode Exit fullscreen mode

The assistant pane in Zed operates as a persistent context window alongside your editor. You can reference files directly in the conversation with @filename, pull in diagnostics, include the current selection, or reference your entire project tree. The context insertion is manual and explicit — you decide what goes in.

That explicitness is actually a feature. With Copilot-style autocomplete you're always guessing what the model has in context. In Zed's assistant pane you know exactly what you sent.

Inline completions work through the same model backend. The latency from a Claude-backed inline completion in Zed is lower than the same request through a VS Code extension because there's no extension host middleware adding overhead.


Claude Code Integration: The Part Nobody Talks About

Claude Code runs in the terminal. Zed's integrated terminal is fast, persistent across sessions, and handles split panes cleanly. That's table stakes. The part that actually changed my workflow is that Zed's file watching and editor state integrate well with Claude Code's filesystem operations.

When Claude Code writes a file, Zed picks it up instantly. No reload prompt, no stale buffer warning, no "file changed on disk" dialog to dismiss. The file just updates. When you're running a Claude Code agent loop that's touching ten files across a project, that frictionless sync is significant.

The workflow I run most:

  1. Zed open on the project
  2. Claude Code in the integrated terminal with claude at the project root
  3. AI assistant pane open for specific questions and code review
  4. Inline completions active for fill-in work

This is three AI surfaces in one editor, with no extensions involved. VS Code can approximate this with Copilot + Claude extension + a terminal session, but you're managing three different context models and three different latency profiles. In Zed it's one coherent system.


The Extension Ecosystem Gap (and Why It Matters Less Than You Think)

Zed does not have VS Code's extension library. That's a real tradeoff. Specific language tooling, niche formatters, and workflow plugins that exist as VS Code extensions may not have Zed equivalents yet.

For the security and embedded work I do, the gap is smaller than expected. Rust support in Zed is excellent — which makes sense given the editor is written in it. Python, TypeScript, Go, and C/C++ language servers work via LSP. Tree-sitter syntax highlighting covers the languages I'm actually in daily.

What I lost: a handful of specific VS Code extensions for CAN DBC file parsing, some custom snippets I'd built up over years (portable to Zed with minor restructuring), and a couple of Git blame visualization plugins. What I kept: everything that matters for the work.

If your workflow is deeply dependent on a specific VS Code extension that has no equivalent, Zed will block you. Audit your actual extension usage before switching. Most people have thirty extensions installed and use six.


Keybindings and the Muscle Memory Problem

Zed supports VS Code keymap imports. The translation is not perfect but it's close enough that the first week is not a complete retraining exercise.

// In settings.json, to use VS Code-style bindings:
{
  "base_keymap": "VSCode"
}

Enter fullscreen mode Exit fullscreen mode

The gaps are in edge cases — specific multi-cursor combinations, panel focus shortcuts, some terminal keybinds. The command palette (Cmd+Shift+P on Mac, Ctrl+Shift+P on Linux) works the same way and covers most of what the keybinding gaps don't.

Give it two weeks before you decide. The first three days are annoying. By week two, the speed of the editor starts compensating for the occasional muscle memory failure.


What VS Code Still Has That Zed Doesn't

Honest accounting:

The extension ecosystem. This is the obvious one. If you need remote SSH development with full IntelliSense, VS Code's Remote Development extension suite has no peer in Zed yet. Zed has remote editing in progress but it's not at feature parity.

Jupyter notebook support. VS Code handles .ipynb files natively. Zed does not have this as of mid-2026. If your workflow lives in notebooks, this is a blocker.

Debugger GUI. VS Code's debugger UI is mature. Zed's debugging support is improving but the VS Code DAP-based debugger with its variable inspection panel is still ahead for most use cases. For embedded work with openocd and GDB, I'm still using VS Code for that specific context.

For everything else in my actual daily workflow, Zed wins on speed and the AI integration coherence makes the tradeoffs worth it.


The Configuration Surface

Zed is configured entirely in JSON. No settings UI maze, no GUI toggles to hunt through. Your settings.json and keymap.json are the full configuration surface.

{
  "theme": "One Dark",
  "buffer_font_family": "Zed Mono",
  "buffer_font_size": 14,
  "vim_mode": true,
  "format_on_save": "on",
  "tab_size": 2,
  "soft_wrap": "editor_width",
  "terminal": {
    "font_family": "Zed Mono",
    "font_size": 13
  }
}

Enter fullscreen mode Exit fullscreen mode

The configuration lives in ~/.config/zed/settings.json on Linux, ~/Library/Application Support/Zed/settings.json on macOS. Dot-file it, version control it, sync it.


Making the Switch

If you're coming from VS Code with an AI-heavy workflow, the friction is lower than you expect. Export your VS Code extensions list first — identify which ones are genuinely critical versus which ones you installed once and forgot about. Map those to Zed equivalents or decide whether you can drop them.

$ code --list-extensions > vscode_extensions.txt

Enter fullscreen mode Exit fullscreen mode

Run Zed alongside VS Code for two weeks rather than hard-switching. Use Zed for new work, VS Code for anything that requires an extension you haven't solved for yet. By the end of two weeks you'll know whether your workflow maps.

The performance difference is real and it compounds. Fast editor, fast AI responses, clean context management. For AI-augmented work specifically, the architecture choice starts to matter in ways it never used to.


If you want to go further with Zed's configuration and power-user workflows, I put together a playbook at numbpilled.gumroad.com/l/zed-playbook — terminal edition, covers keymaps, AI backend configuration, multi-project setups, and the shortcuts worth learning first.

For the Claude Code side of this workflow, the productivity tricks guide is at numbpilled.gumroad.com/l/claude-code-for-devs — 21 patterns that change how the agent loop actually works in practice.