OpenAI Codex now writes your prompts before you do
OpenAI just launched composer predictions in beta for Codex Pro subscribers, guessing what you will type next based on your conversation history and communication style. You can hit Tab to accept the suggestion, edit it, and send. It currently only works for local tasks in the desktop app, but OpenAI calls it one of their most-loved internal features yet.
OpenAI is turning the prompt box into an autocomplete engine.
The company just rolled out composer predictions in beta for Codex Pro users. Instead of waiting for you to type out every follow-up instruction, Codex can now predict what you are about to say next based on your conversation history and personal communication style.
Press Tab to prompt
Right now, the feature is limited to local tasks inside the Codex desktop app. When a suggestion appears, you simply press Tab to accept it, edit the text if needed, and hit send. If you find the AI guessing your next move distracting, you can toggle it off in Settings at any time. OpenAI noted that this was one of the most loved features internally during testing.
Why it matters
Developers already have Tab autocomplete hardwired into their muscle memory for code. Bringing that exact same mechanic to the prompt input itself is a clever shortcut. For power users dealing with repetitive back-and-forth exchanges, like asking to fix a lint error or run tests, shaving off a few keystrokes quickly compounds over a full workday.
Keeping it local to desktop tasks looks like a deliberate choice to prevent latency lag from breaking the flow. The real test is whether the model can accurately mirror an engineer's intent, or if it ends up feeling like just another autocomplete pop-up you routinely ignore.
Sources (1)
- x.com(opens in a new tab)Originalx.com