highsky
December 14, 2024, 7:52am
1
It would be nice if you let me change my question without creating another branch. It’s something like what if function.
How It Could Look in Practice
After receiving a response, you click a “What If” Button beside your original question.
You edit your question, and I recalculate the response dynamically.
The conversation updates seamlessly:
“What If” versions appear inline as variations of the original input.
Original responses remain accessible for comparison (if desired).
1 Like
Hey you might be better off using Macros for most missing features.
I have a Forum Thread introduction and GPT to help to show how to achieve this with a bunch of examples.
Working Name: ‘Phasm’
I have posted a couple of times showing functions I am writing into ChatGPT Plus for some timesaving tasks.
Im()
Create a function called Im() that accepts a single parameter, Description. Save it to Memory and call it whenever referenced in my chats. Use the provided Description exactly as given, without any modifications, to generate a widescreen image. Ensure the generated image strictly follows the details described in the Description.
Something like this allows you…
I hope this helps…
If you want to do this without a GPT directly into Chat GPT you can too but read the article for how to do this with Memory on
WI(I could use Macros)
WI(Scenario) {
1. Retrieve the context or content from the previous message.
2. Combine the provided Scenario with the context of the previous message to form a “What If” question.
3. Formulate the output as:
- “What if [Scenario] in the context of [PreviousMessage]?”
4. If no Scenario is provided, return a clarification or general “What If” based only on the previous message:
- Example: “What if [alternative approach to key concept in PreviousMessage]?”
}
You can see here that ‘Previous Threads’ are saved if you edit (3/3) just click the ‘<’ to go through them
highsky
December 14, 2024, 8:45am
3
I should have said the purpose of what if functions is not to make unnecessary conversation branches.
In which case, the best suggestion I have is to create a custom GPT, ask questions to the What If bit and then create a share link…
Conversation will always start from that point on the share link… Just refresh the page for another WI.
And try something like:
WI(Gravity Worked Backwards)
Though I think this will only work if you have a lot of What Ifs
1 Like
highsky
December 14, 2024, 7:43am
5
It would be great if you let a user be able to freeze a conversation and start a new conversation based on the context of the frozen conversation.
If a conversation gets too long, the web page becomes sluggish.
highsky
December 14, 2024, 7:48am
6
It would be convenient if you let me fold parts of a conversation so that it is not shown on the web page. the context from it should remain, though.
I answered this for you here but check the full context:
Hey you might be better off using Macros for most missing features.
I have a Forum Thread introduction and GPT to help to show how to achieve this with a bunch of examples.
I hope this helps…
If you want to do this without a GPT directly into Chat GPT you can too but read the article for how to do this with Memory on
WI(I could use Macros)
WI(Scenario) {
1. Retrieve the context or content from the previous message.
2. Combine the provided Scenario with the context of the previous mess…
highsky
December 14, 2024, 7:48am
8
It would be convenient if you let me fold parts of a conversation so that it is not shown on the web page. the context from it should remain, though.
highsky
December 16, 2024, 1:05am
9
It would be convenient sometimes when a unique mark is present in users’ questions, you show only your responses not the users’ questions.
User Input with a Mark :
A user includes a predefined mark (e.g., #hide
or ?only
) at the beginning or end of their question.
Example:
Input: “What is the capital of France? #hide ”
Output Behavior :
When you process the input, your response appears without displaying the user’s question .
Example:
Output: “The capital of France is Paris.”
The system could optionally log the input internally for context continuity but keep it invisible on the user interface.
highsky
December 16, 2024, 12:58am
10
It would be nice if you let users prune the entire conversation branches when they think their conversations are complete at that point.
Why This Feature Makes Sense
Focus and Simplicity : Pruning lets users declutter long, branching conversations while keeping them organized.
Better Resource Management : Discarding irrelevant branches reduces the complexity of the conversation, improving response times.
Enhanced User Control : Gives users the power to finalize and tidy up their conversations without losing functionality.
highsky
December 16, 2024, 12:54am
11
I understand you don’t want to let a user to remove branch at will.
Then it would be nice if users were to be able to see the branch tree in a conversation and navigate the tree and go to a certain point.
Why This Feature Matters
It makes navigating and managing complex conversations much easier.
Encourages creative exploration of ideas without fear of losing track.
Empowers users to review and refine questions in a structured way, enhancing the overall experience.