I have an assistant that uses tools to call some endpoints on a private API. It works great, and so I am thinking about optimizations. It would be interesting if I could allow a user to rapid-fire voice commands at the assistant:
“add 1lb of X”, “add 2lb of Y”, “remove 5 of Z”
The Assistant handles this as:( Input > Run > Response) (Input > Run > Response) (Input > Run > Response) in parallel. I’m wondering if there is an advantage in adding multiple consecutive messages with the ‘user’ role in the thread at once. There’s nothing in the documentation indicating this is an intended use of the messages parameter though. Is this not recommended?