What happened: Google says the Gemini app is getting “notebooks,” a feature meant to organize chats and project files, and it will sync notebooks across Gemini and NotebookLM so sources and context carry between the two products.

Why it matters: The real competition is workflow, not vibes. If assistants become useful at scale, they need persistent context, structured sources, and guardrails around what the model should treat as ground truth, which is exactly what “personal knowledge bases” are trying to be.

Wider context: Google frames this as a step toward longer-running projects, with subscription tiers controlling how many sources you can use. That is product strategy meeting compute economics: persistence, context, and limits, sold as “organization.”

Background: The post says notebooks roll out first to Google AI Ultra, Pro, and Plus subscribers on the web, then expand to mobile, more European countries, and free users later. It also notes notebooks won’t be available for under-18s, Workspace, or Education accounts.

Try notebooks in Gemini to easily keep track of projects — Google Blog

Singularity Soup Take: The future is not one perfect chatbot, it’s a thousand slightly less chaotic folders. If Gemini can keep sources, instructions, and threads from turning into a scroll-forever memory leak, that’s an actual feature, not just a new coat of gradient paint.

Key Takeaways:

• Notebooks concept: Google describes notebooks as shared personal knowledge bases where you can group chats, add files like documents and PDFs, and give custom instructions, with the goal of keeping complex projects organized inside the Gemini app.
• Cross-product sync: Notebooks sync between Gemini and NotebookLM, so sources added in one appear in the other. Google highlights using NotebookLM features like video overviews and infographics, then continuing work in Gemini with the same material.
• Access and limits: Rollout starts with Ultra, Pro, and Plus subscribers on the web, with broader access coming later. Availability varies by plan and account type, which effectively turns “how much context you can use” into a pricing lever.

Related News

ChatGPT Search vs the EU DSA: When Your Chatbot Gets Treated Like a Platform - Another example of assistants evolving from “app feature” into something that looks like infrastructure, with rules attached.

Relevant Resources

Understanding ChatGPT and Large Language Models - A friendly explainer on how LLMs use (and sometimes misuse) context and sources.