As an academic author working on extensive projects, I often encounter the token limit in lengthy ChatGPT conversations. While breaking the work into smaller sections is a practical workaround, it is not always efficient. A highly beneficial enhancement would be the ability to delete specific questions and answers within a chat.
This feature would allow users to remove irrelevant or tangential content, effectively managing token usage and maintaining conversation relevance. However, I understand the potential technical and ethical complexities of implementing such functionality. Therefore, I propose that OpenAI introduce this feature with a controlled approach, where users can apply for access. Applications could be reviewed on a case-by-case basis, granting permission only to those with clear, legitimate use cases—such as academic, professional, or research purposes.
This approach would ensure that the feature is used responsibly and transparently while addressing the needs of users working on long-term, in-depth projects. Implementing such a controlled system would greatly enhance the user experience for individuals like myself, who require flexibility and control over extensive ChatGPT interactions.