Hello Open AI forum folks, this is my first post…
I had a brief conversation with GPT. I was looking for a way to use GPT to help with business management and ideas as well as other business related tasks and came up with these two seemingly simple to implement interface upgrades that I assumed were already part of chat GPT (I am not a power user). The simplicity of the implementation are original ideas as far as I can tell, I just used the bot to write it up all nice for me. I saved the conversation that led to these obviously written by AI feature requests if anyone is interested. Here they are:
-
Conversation Retention via Context Refresh Button
Currently, when I return to a previous conversation, I need to manually copy and paste context from past interactions to maintain continuity. I propose adding a button within the chat window that allows users to automatically send past conversations to the system. This feature would refresh the bot’s context, allowing it to seamlessly continue the conversation from where it left off. The button would initiate an automatic “copy and send” process in the background, without cluttering the chat window with repeated text. Instead, only a simple note (e.g., “Conversation context refreshed”) would indicate that past context has been reloaded. Once the button is pressed and the context is updated, it would disappear from the interface to avoid accidentally triggering it again during the same session.
Technical Simplicity: From a technical perspective, this feature would likely be a relatively simple addition. It would primarily involve implementing a button within the chat window and a background process that fetches the past conversation context when the button is pressed. The system would then need to apply that context in real time, which is a relatively straightforward task in terms of user interface (UI) and backend integration. In fact, a simple update could be as minimal as adding a statement like, “Please refer to our previous conversation” as a header for each update request. Emphasizing that no changes whatsoever would need to be made to ChatGPT itself makes this a highly feasible enhancement. -
Integrated, Seamless Feedback Mechanism
To improve user feedback and ensure that suggestions reach OpenAI more directly, I recommend implementing an integrated feedback feature that allows users to submit ideas or feedback within the conversation itself. This mechanism could automatically capture and send feedback during chats, eliminating the need to navigate through the app’s settings or feedback forms. Users could submit their thoughts without interrupting the conversation flow, and OpenAI could receive suggestions in real time to enhance the system’s responsiveness to user needs.
Technical Simplicity: This update could be implemented entirely within the app interface, without requiring any modifications to ChatGPT’s core functionality. The feedback system could use an AI model trained to recognize when a user is suggesting a new feature or making a direct request for the bot to pass along feedback. This AI would detect keywords or patterns related to feedback or suggestions, triggering an automatic message to OpenAI. For example, if a user mentions, “Can you suggest this feature to OpenAI?” the AI could recognize the request and automatically send the feedback. Initially, this feature could be integrated into the interface using a simple rules-based AI that identifies these requests. Later, as the system evolves, it could be upgraded to a more advanced model directly integrated with ChatGPT, allowing for higher accuracy and more nuanced feedback identification. This would be an enhancement that doesn’t require changes to the core chatbot programming but instead leverages interface-level AI detection and automated actions. Eventually, it could become an entirely seamless feature with minimal user input.
I would use both of these features regularly, any of you think you could benefit from them?
As an afterthought, the system of resending conversations will also allow for GPT to correct past errors in the conversation (such as math errors) as it progresses in accuracy.
Cheers
Josh