Feedback on the Plan Mode Feature: Initial Impressions and Suggestions for Improvement

hey Daphné, here is my first feedback on the feature “plan mode” you made available for me last week. I used it for the first time today. I was working on a plan for a conference based on a transcription from a brief call with my client + notes in an email + context in my Notion. So I threw everything at my Dust agent, asking it to use plan mode to structure our work in order to clarify my thoughts before having a plan for the conference and a proposal email for my client. The first result was a plan on the side panel, pretty good + a very loooong message in the chat, whereas I don’t think the task was this complicated. I think maybe it’s because of ChatGPT Luna, which is bad, btw, but I forgot to change to Claude. -> first feedback here: I don’t see the point to have the results in the .md file + in the chat. Anyway, the plan was good, so I went along with it. We worked for a bit, the agent crossed off the tasks on its own, which is cool. But finally, I edited the plan of my conference and asked it to remove the old one from the conv history -> it deleted the plan made by plan mode because it made the confusion between the plan (from our work together) and the plan for my conference. I understand why it did the confusion, but still think it’s kind of numb. Also, I don’t really see the added value of having this plan mode on the side. Although I do ask my agent on a regular basic to explain to me the plan it is going to follow in order to achieve the goal I set up. So, I think it’s useful to have a plan mode. But I’m not convinced yet with the feature. Maybe I need to use it a bit more. Will keep you posted on my next try.