Google updates NotebookLM expanding analysis capacity and role customization

Google has introduced updates to NotebookLM that substantially modify the capacity of its artificial intelligence-based chat system. The main improvement expands the context window to one million tokens using Gemini models, allowing the processing of complete document collections without size limitations.

According to Google, internal testing indicates a 50% increase in user satisfaction when working with extensive source document sets. Additionally, the system has expanded the length of active conversations by more than six times, allowing much longer exchanges without losing coherence.

NotebookLM has modified its information search and synthesis method. The system now automatically explores sources from multiple perspectives, generating responses that incorporate different approaches present in the consulted sources. This process is especially relevant in notebooks with large volumes of documentation, where the selection of appropriate context determines response accuracy.

The tool incorporates automatic saving of conversation history, allowing users to close and resume sessions without losing the record of previous interactions. Users can delete this history at any time, and in shared notebooks, each conversation remains visible only to whoever generated it. This functionality is being progressively distributed during the current week.

Chat customization now allows defining specific goals, voices, or roles for the assistant. Users can configure behavior through instructions that establish, for example, that the system acts as an academic advisor, marketing strategist, or analyst from academic, creative, and critical perspectives. Configuration is performed from the chat interface, where the user can define the desired response style.

Google indicates these modifications seek to adapt NotebookLM to different types of projects through more precise and contextualized responses. The expanded context window is available in all tool plans, while role customization options are accessible to all NotebookLM users.

Key points

• NotebookLM expands its context window to one million tokens with Gemini models to process large document volumes
• Allows customizing assistant roles, such as research advisor or marketing strategist, to adapt responses to specific goals
• The system automatically explores sources from multiple perspectives, integrating detailed analyses in responses
• Expands multi-turn conversation capacity by more than six times, maintaining coherence in prolonged interactions
• Incorporates automatic chat history saving, with the option to delete it at any time
• Google's internal testing reports a 50% increase in satisfaction when handling extensive document collections
• In shared notebooks, conversation history is visible only to the user who generated it