Key Takeaways
- OpenAI merged its dedicated ChatGPT desktop app with the existing Codex app, creating a unified “super app” experience.
- This move forced users of the previously native ChatGPT app onto an Electron-based framework, which relies on web views instead of direct system integrations.
- The shift sparked significant user backlash, citing a degraded experience, loss of familiar shortcuts, and confusion around new features.
- OpenAI's Tibo publicly acknowledged the issues, admitting they “didn't get everything quite right,” specifically regarding compute settings, app reorganization, and unclear usage limit messaging.
The “Super App” That Broke the Experience
OpenAI, a company known for pushing the limits of AI, just learned a tough lesson about user experience. They rolled out a new ChatGPT desktop app, but it wasn't really new. As John Coogan explained, “The new chatbT app is just the Codex app.” This unification into a single "super app" meant integrating ChatGPT's features into the existing Codex application. The problem? The Codex app was built with Electron, a framework that essentially wraps web views into a desktop application. The original ChatGPT desktop app, however, was a native application, offering tighter system integrations and a snappier feel.
This technical choice triggered an immediate backlash from users. “A lot of people don't like that because the Codeex app is written in Electron which means it's a bunch of like web views, web browsers effectively as opposed to uh the chat app which was a native app,” Coogan noted. Users who had grown accustomed to the native app's speed and its deep connections with their operating system found the new Electron version clunky, slower, and lacking the familiar shortcuts and integrations they'd come to rely on. It was a classic case of internal engineering convenience clashing hard with established user expectation.
Even OpenAI Can Miss the Mark
Within 24 hours of the rollout, the user revolt was loud enough to get OpenAI's attention. Tibo, an OpenAI representative, quickly addressed the feedback. “We spent the last 24 hours reading feedback, looking at usage patterns and talking with many of you,” he shared. The short version: there was excitement for the new 5.6 update, but “we also didn't get everything quite right.” This candid admission pointed to several critical missteps.
Specifically, Tibo highlighted two key areas of failure. First, OpenAI “made it too easy to use the highest compute settings without making the impact on usage limits sufficiently clear.” Users unknowingly burned through their limits faster. Second, they “reorganized the desktop app in one bold move, making familiar things like chats and projects harder to find.” This wasn't just a technical swap; it was a wholesale reorganization that disrupted core user workflows without adequate preparation or communication. Even a company like OpenAI, with vast resources and a brilliant engineering team, can stumble when they move fast and break things without deeply understanding the user impact of technical decisions.
What to Do With This
If you're building a product, this is a stark reminder: don't sacrifice core user experience for internal efficiency or a grand "super app" vision, especially when replacing something your users already love. Before a major technical re-platforming—like switching from native to Electron—identify your 100 most active, power users. Map their absolute critical workflows, their favored shortcuts, and their performance expectations. Launch an early access program with only these users for a full month. If even one key workflow is slower, clunkier, or broken, halt the public release and redesign. Never assume your internal metrics reflect the raw, unfiltered frustration of your external user base.