One of ChatGPT's most practical features just disappeared without so much as a changelog entry. Users can no longer click on a previous message in a conversation thread and edit it in place - a capability that let you tweak a prompt, fix a typo, or try a different angle without starting over from scratch.
Previously, clicking on any of your earlier messages would let you modify the text and regenerate the response from that point forward, preserving the conversation history above it. Now, your options are limited to regenerating the entire response or starting a new branch. For anyone who iterates heavily on prompts - writers refining output, developers debugging code suggestions, marketers testing different angles - this adds real friction to the workflow. Instead of surgically editing one message, you are stuck copying your text, starting fresh, and losing the context that built up over multiple exchanges.
OpenAI has not announced the change or explained the reasoning behind it. The removal appears to be rolling out gradually, with some users still seeing the old behavior while others have lost access. This follows a pattern where OpenAI ships interface changes without prior notice, which has become a recurring frustration among power users.
The timing is particularly odd given that competing tools are moving in the opposite direction. Claude lets you edit and retry any message in a conversation. Google's Gemini supports similar editing. Removing a feature your competitors actively promote is a strange move, especially one that costs nothing to maintain and saves users significant time.
If the change sticks, expect workarounds to emerge - browser extensions, custom instructions to minimize re-prompting, or simply more users exploring alternatives that still offer granular conversation control.