I have a specific request regarding code responses:
Users should be able to edit code output directly within the chat interface to correct errors. Instead of repeatedly displaying the same script, there would be one script that both ChatGPT and the user can modify. This functionality resembles “Supervised Learning,” though I’m not an expert.
These updates wouldn’t alter the underlying model of the language model. They might involve creating a new model, but that’s not essential.
A suggested implementation: After a user edits the code, the next prompt to the ChatGPT engine could acknowledge the changes and continue the conversation seamlessly, without appending new scripts or using animations to highlight edits.
Thank you for considering this enhancement.