Google Gemini: Ideas and Opinions — Chapter #02

In March 2025, we published an article on this project with ideas and opinions for enhancing Google Gemini. Here, we are publishing our second article with ideas and recommendations aimed at improving Google Gemini's over-all usability.

Table of Contents

1. Enable Prompt Copying
2. Inline Response Mode
3. Response Convergence Point or Format Lock-in
4. See also

① Enable Prompt Copying (Browser)

In our previous article, we wrote about the absence of an easy way to reuse a prompt. Other than the "Reuse" option, it is also important for users to have an easy option to copy their own prompts. At present, AI responses might not be entirely accurate always, and users might need to ask the same or similar questions repeatedly. Additionally, users may want to reuse a particular question in the same or in different thread (or outside Gemini). At present, there is not an easy way to copy and paste a user's prompt. On Google Gemini, a lengthy prompt is displayed in a collapsed box, making selection difficult. This makes the copying more cumbersome. Hence, implementing a "copy" option for user prompts, similar to the feature available for code blocks and documents in the AI's responses, would be very useful.
Note: On Gemini mobile app, the prompt copy option is available.

② Inline Response Mode

Google Gemini, similar to other generative AI models, follows an instant messaging (chat) format. This would be useful to incorporate an inline response feature, similar to how one can select and respond to a specific message in instant messaging applications. This inline response option would be particularly beneficial in two scenarios:

1. Clarity in lengthy threads: A conversation thread might contain numerous messages, and in such scenario, users could quickly highlight a specific message to clearly point out the subject of their next query or the message they are referencing.
2. Targeted modifications in long responses: When Google Gemini gives a lengthy answer and a user wants to suggest a modification to a specific section, they could quickly highlight a particular portion, and ask their questions around it.

③ Response Convergence Point or Format Lock-in

This is a common experience when interacting with Google Gemini that achieving the desired output format and structure can be a time-taking iterative process. Let us consider the following interaction:

1. User asks a question.

2. Google Gemini gives an initial response.

3. User suggests a specific modification or change.

4. Google Gemini revises the answer accordingly but introduces another error.

5. The user highlights the new error and requests a further correction.

6. Google Gemini makes that requested correction.

This cycle of response and refinement might take place several times. Finally, Google Gemini produces an answer that perfectly aligns with the user's specified format, structure, and requirements. We can identify this point as the "Response Convergence Point."

At this point of convergence, a helpful next step would be to "lock in" this desired format. We have noted that explicitly asking Gemini to "follow this exact format in future answers" doesn't ensure consistent adherence. Gemini still deviate quite often, necessitating the entire refinement process to begin from the scratch.

We need "inline response", "convergence point". 

To reduce repetitive effort, especially when someone is working with large datasets and similar queries, the ability to lock in this "Response Convergence Point" should be beneficial. Possibly, users might have a way to tag and label a satisfactory response format for Gemini's and their own future reference. This would significantly streamline workflows and ensure consistency in subsequent questions and answers.

See also

• Google Gemini — Ideas and Opinions #01, the first chapter of this series