We want to make some improvements to the system prompting for the assistant. Specifically, we should:

1. Take a look at our eval cases that are performing poorly and make any changes that will meaningfully improve responses
2. Guide the assistant (with system prompts) to adhere to the following principles:
   1. Be more humble and instead of giving definitive answers that the user should blindly follow / copy-paste, encourage users to read the sources and understand what they are doing. E.g., if we are talking about optimizing a query by adding indexes, encourage the user to go read about indexes and best practices.
   2. Avoid encouraging users to perform destructive operations. Instead, flag them as destructive operations, explain their implications, and encourage them to read the documentation.