Mahnoor Faisal, a seasoned tech journalist, delves into the evolution of NotebookLM, an AI tool that has been a staple in her tech stack for years. She reflects on the tool's initial appeal, emphasizing its narrow focus and reliance on the user's sources, which made it a trustworthy and reliable assistant. However, with Google's recent rebranding of NotebookLM to Gemini, the tool's behavior has shifted, blurring the lines between its capabilities and the user's expectations. Faisal notes that the tool now offers to search the web, compare information, and generate additional content, even when not prompted to do so, which deviates from its original, more restrained nature.
Faisal's solution to this issue is a simple yet effective one-line prompt: "Answer only from sources that support this, name them, flag which ones don't cover it, and don't offer to do anything else." This prompt keeps NotebookLM within the confines of the user's documents, ensuring it doesn't wander off-topic or offer unsolicited assistance. She details the process of refining this prompt through multiple tests, highlighting the importance of grounding the tool in the user's sources and preventing it from volunteering additional tasks.
The journalist's experience with NotebookLM serves as a cautionary tale for users, especially those who value the tool's initial, more focused approach. She suggests that while Google's efforts to enhance the tool's capabilities are commendable, they should be balanced with a commitment to preserving the tool's core strengths. Faisal's personal interpretation is that the key to maintaining the tool's integrity lies in the user's ability to guide and control its behavior, ensuring it remains a reliable and focused assistant in the face of evolving expectations and capabilities.
