NotebookLM
Gemini
Google
TLDR: Google upgraded NotebookLM with Gemini 3.5 and Antigravity plus a secure cloud computer, improving analysis and adding download formats. It also expands source discovery and edits.
Key Takeaways:
- Google rolled out NotebookLM upgrades for AI Ultra and Workspace business customers, expanding models, research workflows, and output options.
- NotebookLM now uses Gemini 3.5 and Antigravity and includes over 100 curated software skills plus secure cloud code execution.
- Side by side evaluations show win rates above 65%, 69.9%, and 78.2%, while outputs download as PNG, SVG, PDF, DOCX, XLSX, and PPTX.
Google is turning NotebookLM from a helpful assistant into a small research workshop, complete with code execution and exportable deliverables. The big question now is whether users will trust the outputs as much as they trust spreadsheets and slide decks.
Google is turning NotebookLM from a helpful assistant into a small research workshop, complete with code execution and exportable deliverables. The big question now is whether users will trust the outputs as much as they trust spreadsheets and slide decks.
Q&A
What changes when NotebookLM can write and run code inside a secure cloud computer?
Users can push analysis from “summarize” to “compute,” reducing manual glue work between research notes, data cleaning, and chart production.
Why do file formats matter more than model upgrades for day to day work?
If exports land in PDF, DOCX, XLSX, and PPTX directly, teams spend less time converting outputs, which makes the tool feel faster and more trustworthy.
How might the win rate improvements shift what researchers ask NotebookLM to do?
Higher source discovery and large document analysis scores encourage more ambitious prompts that mix citation hunting, cross language search, and synthesis.
What could the “build your source repository” chat capability change in real projects?
Teams may start projects earlier by assembling a living evidence set, which can shorten the usual loop of collecting sources before analysis begins.
What happens as Google expands these features beyond AI Ultra and Workspace business tiers?
Wider access could accelerate competition for document intelligence tools, but it will also test whether performance remains consistent across different usage levels and permissions.
No comments yet. Be the first to share your thoughts!