Going Native

Going native

Desktop apps defaulted to Electron for a decade. It solved a real problem: write once, ship everywhere, reuse web skills you already have. AI coding assistants changed the math. Writing native desktop software costs a fraction of what it used to, and AI models do their best work on paths that have been walked before.

Modern native frameworks are built on exactly that: less indirection, fewer hidden abstractions, a shorter trip to the metal.

Rust and GPUI

GPUI is the framework at the center of this project, built by the team behind Zed. A GPU-accelerated UI framework for Rust. If you've used React you already get the model: elements and state, a reactive render loop, all running at native speed.

Rust gives you memory safety without a garbage collector. Fearless concurrency. A type system that catches bugs before they ship. Combine that with modern AI models and the learning curve isn't what it was.

The ecosystem is young: video, rich text, platform integrations. The community is building them out. You're coding against the platform directly, nothing between you and the OS.

Why it matters

Every Electron app ships its own copy of Chromium. It doesn't matter if it's a chat client, a notes app, or a settings panel. Each one hauls a browser engine. Native apps sip memory, launch fast enough that you never see a splash screen, and fit the machine they run on.

Pick the right tool for the work. The barriers dissolved. AI handles most of the implementation at this point, whichever stack you choose. A web stack stuffed into a desktop app is overhead you can drop.