Google CEO Promotes Major Upgrade to NotebookLM
Google has upgraded the underlying model of NotebookLM to Gemini 3.5 and integrated Antigravity.
Each Notebook features a built-in secure cloud computer, supports writing and running code, and comes pre-installed with over 100 software skills for complex research and data analysis.
New features include more than ten output formats, such as PDF, docx, Excel, PPT, charts, and images, which can be edited after generation, and integrates with Google Search to initiate projects without needing to bring in materials.
Source: Public Information
ABAB AI Insight
Google has previously enhanced its research tool positioning through iterations of NotebookLM, including the introduction of Audio Overview and multimodal source support. The integration of Gemini 3.5 and Antigravity continues its path of evolving consumer tools into agentic workstations, having previously reduced hallucination issues through source citation mechanisms in earlier versions.
On the capital path, Google mobilizes resources from the Gemini model family and the internal Antigravity coding environment, embedding them directly into the NotebookLM cloud execution environment. This move lowers the local computing threshold for users and forms a paid closed loop through AI Ultra and Workspace enterprise subscriptions, accelerating the transition from a free research assistant to a high-value productivity platform.
Similar to how OpenAI transitioned from accumulating users with ChatGPT to advanced subscriptions and toolchain integration, NotebookLM is currently in a transformation phase from a note-taking tool to a full-stack AI research workstation, allowing Google to strengthen its control in the enterprise knowledge work domain.
Essentially, this represents a technological replacement and industrial chain restructuring: Gemini 3.5 combined with Antigravity allows NotebookLM to directly replace traditional local IDEs and multi-software collaboration links, compressing research cycles through cloud-based execution and search integration, forcing professional toolchains to migrate to the Google ecosystem and reshaping the division of knowledge production.
ABAB News · Cognitive Law
The smarter the tool, the more users rely on structure rather than skills.
Free entry accumulates into a tower, while paid closed loops forge a moat.
Agent execution replaces local operations, shifting control from users to the platform.