app Notes
input I’m not quite sure we need the decision and request idea concept I think it can all be stored in a decision. Decision can store the user assistant response and the bug and thinking process and even if I’m a synchronous, this will create or a new session or reuse an existing one and the feedback would get updated in this session so if I want to know where we’re at or what’s happening I just have to check session.
In memory module when creating new memory, we don’t need to pop up the model window like we do. We can just open up a new window that is gonna be always associated to adding memories
You need to be able for the status message to be separated from the current feedback message or no current feedback stay the same but we added a new one that is given answer and the feedback message. The concept is that we post the first one but after that, we exit it to add to it so it would be kind of loading skid XXX, executing comment, and after that the answer on a new message
When I think a memory also give me a call a session ID or a memory ID that I will open up the detail memory from in my iOS
And is Dev environment variable to say that we’re in development and the difference would be to load flask versus Uvicorn
Remove the concept of dynamic system prompt with two usage in the agent and the system. Proulx should be listing the tool that we have but should not change over time and to get loaded and introduce Miss rotation so if message get over a limit, we will remove the oldest message and if write the answer to that message is a tool usage, but we remove the tool as well I think
I want to come out all the code related to summarizing and don’t want to use any summary in the web user interface. The summary was something that was planned for when using vocal agent, but we are not doing that anymore. Now we are only doing written feedback. I still wanna keep the summarize functionality the general function, but I don’t want it to be called or any summary to be stored anywhere.
I want you to run extended test on the time it take to process a webpage using Charlotte versus using image from it. I want to compare which one is faster so do a full comprehensive test of long webpage and give me the results
I want to add the concept of users in the agent app
Add the concept of device