Claude Code + NoteBookLM = Infinite Memory
The video demonstrates how to integrate Google's NotebookLM with Claude (both Code and Co-work versions) to create a persistent memory system. This combination allows Claude to remember conversations across sessions without token costs while unlocking content creation and research capabilities through NotebookLM's AI features.
Summary
Jack Roberts presents a comprehensive tutorial on connecting Claude AI with Google's NotebookLM to solve Claude's memory limitations. He explains that Claude suffers from 'amnesia' - starting each session from zero and forgetting previous conversations, which leads to increased token costs when users repeatedly upload context files. The integration involves two main skills: a Claude NotebookLM skill for initial setup and authentication, and a wrap-up skill for storing conversation data. The setup process requires downloading Python scripts from GitHub, authenticating with NotebookLM through browser cookies, and installing skills in both Claude Code and Claude Co-work environments. Roberts demonstrates three key use cases: enrichment (taking existing projects and generating deeper insights using NotebookLM's research capabilities), content multiplication (creating infographics, videos, and other assets programmatically), and long-term memory management through a 'wrap-up' system. The wrap-up feature stores entire conversation sessions in a personal 'brain' notebook that can be semantically searched later. This approach enables users to access vast amounts of historical data through single API calls rather than maintaining expensive context windows. NotebookLM's capabilities include deep research synthesis, competitive intelligence, market analysis, and various content generation features - all without additional token costs since processing happens on Google's servers.
Key Insights
- Roberts argues that Claude has 'amnesia' where every session starts from zero and it forgets things easily, with users often running out of tokens quickly due to repeatedly uploading context files
- Roberts claims that NotebookLM functions as a RAG system that allows grabbing specific needed information instead of cramming several hundred pages of context about business or meetings into Claude
- Roberts demonstrates that NotebookLM can now create cinematic videos automatically, showing an example video about 'the mathematics of agentic drift' generated by the platform
- Roberts explains that the wrap-up skill creates an ever-growing system in NotebookLM that can retrieve any information with one simple call, keeping all data in one place without building up an insane context window in Claude
- Roberts argues that using NotebookLM for semantic search is superior to processing massive amounts of information simultaneously, comparing it to having millions of books where you only retrieve the specific things you need
Topics
Transcript
[0:00] Imagine combining Claude with the world's number one intelligence platform, Google's Notebook LM. I did this and it solves Claude's number one problem, giving Claude a long-term memory that will cut your costs and a lot of capabilities that 99% of people don't even know exists yet. And in this video, I'll show you the exact setup so you can get this full system with near infinite memory, with virtually zero [music] token cost, with a setup that just takes a few minutes. And if you don't know who I am, my name is Jack [0:30] Roberts. I built and sold my last tech startup [music] with over 60,000 customers. And now I teach thousand entrepreneurs and run AI…
Full transcript available for MurmurCast members
Sign Up to AccessMore from Jack Roberts
100 hours of Hermes Agent lessons in 23 minutes
Jack walks through advanced features of Hermes Agent, an AI personal assistant, covering memory systems, background tasks, scheduled cron jobs, model switching, and integration with tools like Obsidian, GitHub, and various AI models. The video aims to help users unlock capabilities beyond basic chatbot usage. Key themes include connecting external memory systems, delegating tasks to specialized AI models, and building a persistent, context-aware personal assistant.
Claude Code = $10,000 Beautiful Slides
Jack demonstrates a GitHub-based system using Claude Code to generate professional slide decks from any website URL by codifying 20 universal design principles. The system uses Firecrawl to extract brand DNA from websites and can integrate AI-generated images via APIs like Krea.ai. The result is polished, brand-consistent HTML presentations created in minutes with minimal input.
The most valuable thing I got from working at McDonald's
The speaker reflects on their first job at McDonald's at age 17, earning £5 an hour. The most valuable takeaway was the realization that they never wanted to work such a job again, reinforced by calculating how long it would take to earn a million pounds at that wage.
Claude Code Memory System = CHEAT CODE
Jack Roberts presents a three-tier Claude memory system designed to give AI tools persistent context across all platforms and sessions. The system consists of short-term identity memory, mid-term project memory via structured folders and claude.md files, and long-term memory using either Obsidian or Pinecone for archiving conversations and expert knowledge. The goal is to eliminate the 'amnesia' problem where AI loses context between chats.
ChatGPT ads aren't actually a bad thing. Here's when I want them.
The speaker argues that ads in ChatGPT are not inherently bad, particularly when users are actively seeking product recommendations. They distinguish this as 'pull advertising,' where the user initiates the request and ChatGPT synthesizes results based on provided context.