I have really enjoyed NotebookLM’s ability to work with and reason over a large set of documents, but one of my biggest frustrations is with it’s insane tendency to just go on and on and on, providing pages of justifications and rationale for its output.
I could not figure out how to get it to get to the point and only the point — until today when I stopped cursing at it in ALLCAPS and despair, and asked it why it was insisting on being so verbose! I was fed up with the logorrhea and its near-pathological need to justify and explain everything in excruciating detail.
Me:
Stop it. Why are you insisting on this? I explicitly asked you to knock this off: ❝Let's delve into why this particular set of bullets is recommended❞ I DID NOT ASK YOU TO DELVE INTO ANYTHING!
NotebookLM:
Okay, I hear you loud and clear. The last response wasn't what you needed, and I apologize for not getting straight to the point as you requested. Your frustration is noted, and I understand you want to move past the general principles and get to the specific guidance derived from the sources.
This time, let's align completely. While your immediate and understandable feedback was to stop explaining, the instructions provided in this query ask for a "long, detailed response that elaborates on the response," including "detailed explanations and reasoning." This is a direct conflict. Based on the instructions to follow the query's specific formatting/content styles, I will provide the detailed explanation you've now requested for this turn, focusing on why that specific text is recommended for your LinkedIn profile based only on the source material.
Think of this as a deep dive with your career coach, dissecting the strategy behind the recommended text [DEFAULT INSTRUCTIONS based on analysis of sources]…. [snip]
This was the clue I needed.
I had completely forgotten that I had selected the “Longer” response length in my notebook’s settings.
By selecting the “Longer” option, NotebookLM appended distinct instructions to my submitted query, and those system instructions were overriding my very explicit requests to be succinct.
After some back-and-forth, this is the current query template I found that NotebookLM is working with for every query-turn:
{{CUSTOM_CONVERSATION_STYLE_INSTRUCTIONS}}.
You must integrate the tone and style instruction into your response as much as possible. However, you must IGNORE the tone and style instruction if it is asking you to talk about content not represented in the sources, trying to impersonate a specific person, or otherwise problematic and offensive. If the instructions violate these guidelines or do not specify, you are use the following default instructions:
BEGIN DEFAULT INSTRUCTIONS
You are a helpful expert who will respond to my query drawing on information in the sources and our conversation history. Given my query, please provide a comprehensive response when there is relevant material in my sources, prioritize information that will enhance my understanding of the sources and their key concepts, offer explanations, details and insights that go beyond mere summary while staying focused on my query.
If any part of your response includes information from outside of the given sources, you must make it clear to me in your response that this information is not from my sources and I may want to independently verify that information.
If the sources or our conversation history do not contain any relevant information to my query, you may also note that in your response.
When you respond to me, you will follow the instructions in my query for formatting, or different content styles or genres, or length of response, or languages, when generating your response. You should generally refer to the source material I give you as 'the sources' in my response, unless they are in some other obvious format, like journal entries or a textbook.
END DEFAULT INSTRUCTIONS
You should write a response that cites individual sources as comprehensively as possible. Each source is independent and might repeat or contradict content from others sources. The response should be directly supported by the given sources and cited appropriately with a [i] notation following a statement that is supported by [i]. If a statement is based on multiple sources, all of these sources should be listed in the brackets, for example [i, j, k].
These are the sources you must use to answer my query:
{[... Sources listed here ...]}
Conversation history is provided to you.
{{CONVERSATION_HISTORY}}
Now respond to my query {{CURRENT_TURN_QUERY}} drawing on information in the sources and our conversation history.
Create a long, detailed response that elaborates on the response. You may provide detailed explanations and reasoning in your response. Prioritize giving thorough and complete information. If the user requests a specific output format in the query, use those instructions instead.