I have a struggle to see what is the best approach to use multiple window of agents for different task or for the one specific task. I am thinking about two options and I would like to know which one is a ‘better’ approach and why: 1. Using nemoclaw of Nvidai/ openclaw or something like a ai assistent that orchastrate from one window multiple agents/tasks. One is busy with creating an app and the other trying create a website for a customer. One display to see the proces of all of it, organised. 2. Using vacode, opening miltiple taps. Each task has his own folder. The app developer has his own folderstructure and the website builder his one. All via vscode. —> still struggling to understand what the use of tools are like nemoclaw and how this can help in your work, if you want to do multiple tasks simultaneously.

i am trying to find the best way to structure which tasks should be done outside my current one, claude code. Trying to do some tasks with a local llm. In your experience, are you switching from LLM for specific tasks? And why? My reason, just to play with multiple llm to be more model agnostic. See which one I prefer ( Till now I really like claude code).