Discord is dropping Electron and rebuilding its desktop app in Rust with GPUI. The news hit the Hacker News front page on November 7, and the replies split exactly the way you'd expect: one camp cheering, one camp saying "good luck rewriting everything," and a third camp quietly checking how big their own Electron app is.

The number that stuck with me: the Electron build of Discord reportedly sits around 290MB with two accounts signed in. Two hundred and ninety megabytes. For a chat app. An app whose core job is showing text messages and playing voice.

I keep thinking about that number, because it makes the whole argument in one figure. Desktop bloat is not a law of nature. It's a choice — one the industry made a decade ago, and one Discord has now decided to unmake.

Electron made sense in 2016. That's the honest part.

Credit where it's due, because the "Electron is trash" crowd likes to forget this: Electron let small teams ship a real desktop app with the people they already had. One codebase, web tech, Windows and Mac on day one. No hiring a C++ team. No maintaining three separate apps. For a startup trying to get a chat app into people's hands fast, that trade was obviously correct. Discord, Slack, VS Code — half the apps on your dock exist because Electron made them cheap to build.

And VS Code is the example Electron's defenders always reach for, because it's the best one. A fast, capable editor that happens to be Electron. It proves the framework isn't inherently slow. A good team can make it sing.

But "a good team can make it sing" is doing a lot of work in that sentence. For every VS Code there's a dozen apps that ship a whole Chromium browser to render a settings page. The 290MB figure isn't a bug in Electron. It's what happens when the easy path is also the heavy path, and nobody ever has a reason to take the hard one.

The bill came due

Here's the thing about trade-offs: they don't stay still. In 2016, shipping fast mattered more than anything, and a 290MB download was an acceptable price. In 2026, users have spent a decade watching chat apps eat RAM like it's free, and the patience is gone. Laptops haven't gotten 290MB more forgiving. The download is still 290MB. The startup is still slow. The memory usage is still the first thing people complain about in reviews.

Discord's move says the maths flipped. The cost of maintaining a native codebase in Rust is now lower than the cost of shipping a bloated app to millions of people who notice. That's not a technical argument, it's a business one — and it's the kind that actually moves companies. Nobody rewrites a working app for fun. They rewrite it when the old stack starts costing more than the rewrite.

Rust plus GPUI is a bet that the native route is viable again for teams that aren't Apple or Microsoft. GPUI is the UI framework behind Zed, the editor, and it's built for exactly this: fast, native rendering without dragging a browser along. If Discord pulls it off, every Electron app with a big user base is going to get the same question from its own team: what's our excuse?

So what does this mean if you ship desktop software?

Probably nothing, tomorrow. Electron isn't dying. There are thousands of Electron apps, most of them fine, most of their users happy enough. Discord leaving doesn't make your app worse.

But it does move the goalposts. "Everyone uses Electron" stops being a defence. When the biggest chat app on the planet decides the framework costs too much, the default answer to "why is our app 290MB?" gets harder. Not impossible — VS Code still exists, the one-codebase argument still exists — just harder. You need a better reason than "that's what everyone does."

And for users, it's good news regardless of whether the rewrite succeeds. Competition between approaches is what keeps software from settling into comfortable bloat. Electron got complacent because it won. A high-profile defection is the market working the way it's supposed to.

My prediction, for what it's worth: Discord's new app will be faster and smaller, the rewrite will take longer and hurt more than anyone admits publicly, and in two years some other big app will cite Discord as the reason it did the same thing. The 290MB download was never inevitable. It was a choice, made in 2016, that made sense at the time. The honest thing is admitting the time changed — and Discord just did.