Google's NotebookLM Gets Its Biggest Update Yet—and Reviewers Are Still Smitten
A fresh CNET review and hands-on reports confirm NotebookLM remains a uniquely practical AI tool after its largest update to date.
What matters
- CNET published a positive review of Google's NotebookLM, calling it a uniquely practical AI tool for working with your own sources.
- Google recently shipped what it calls NotebookLM's biggest update yet, adding visible reasoning so users can see the model's intermediate steps.
- NotebookLM was announced at Google I/O 2023 as an experimental AI-first notebook powered by Gemini.
- Longtime users report the update makes the tool feel less like its original focused self, for better and worse.
- Audio Overviews remain a standout feature, generating podcast-style summaries from user-supplied documents.
What happened
CNET published a full review of Google's NotebookLM on June 24, 2026, declaring it an AI tool "unlike any other" and praising its ability to transform information in surprising ways. NotebookLM, originally announced at Google I/O 2023 as an experimental AI-first notebook and powered by Gemini, lets users bring their own sources—documents, links, notes—and query them with grounded AI assistance.
The review lands shortly after what Google itself describes as NotebookLM's biggest update yet. According to XDA Developers, the update introduces visible reasoning: NotebookLM now shows you what it's "thinking" before it answers, letting users expand and inspect the model's intermediate steps. This is a meaningful shift for a tool that, as XDA notes, had "remained largely unchanged at its core for a long while."
Notably, Google barely mentioned NotebookLM at I/O 2026, leading some observers to fear the tool might be quietly shelved—a familiar pattern for Google products. Instead, the company delivered this major update days later.
Why it matters
NotebookLM occupies an unusual niche in the AI landscape. Rather than functioning as a general-purpose chatbot, it grounds its responses in user-supplied sources, making it a practical research and learning companion. Android Police reports that after nearly a year of use, the tool has become indispensable—not just for fast learning, but for capabilities that go beyond what Google has traditionally highlighted. Audio Overviews, which generate podcast-style summaries from source material, are singled out as a standout feature.
The new visible-reasoning feature matters because it addresses a core trust problem with AI tools: users want to see how conclusions are reached, especially when those conclusions are drawn from their own documents. By exposing its work, NotebookLM is moving toward greater transparency.
However, XDA's hands-on suggests the update is a double-edged sword, making NotebookLM "feel less like itself (for better and worse)." The implication is that as the tool gains mainstream AI features, it risks losing the focused, source-grounded simplicity that made it distinctive.
Public reaction
No strong public signal was available from Reddit or other discussion forums at the time of writing. The conversation so far is dominated by editorial reviews and hands-on reports rather than broad community discussion.
What to watch
- Whether the visible-reasoning feature improves trust and adoption or adds friction for casual users.
- Whether Google continues to invest in NotebookLM or lets it drift, given its near-absence from I/O 2026.
- How the tool evolves as it gains more general-purpose AI capabilities—whether it retains its source-grounded identity or converges with standard Gemini.
- Whether Audio Overviews and other learning-focused features expand to new content types and languages.
Sources
Public reaction
No Reddit or public discussion threads were captured for this story at the time of writing. The available signal comes from editorial reviews at CNET, XDA Developers, and Android Police rather than community forums.
Signals
- Editorial enthusiasm for NotebookLM's source-grounded approach
- Concern from longtime users that the latest update changes the tool's identity
- Anxiety about Google's commitment after NotebookLM's near-absence from I/O 2026
Open questions
- Will the visible-reasoning feature improve user trust or add unwanted complexity?
- Is Google committed to NotebookLM long-term, or will it follow the path of other shelved experiments?
- Will the tool's source-grounded identity survive as it gains more general AI features?
What to do next
Developers
Try NotebookLM's new visible-reasoning feature with a multi-document source set and evaluate whether the exposed reasoning steps improve verifiability of answers.
The update's transparency feature is directly relevant to building trust in grounded AI applications.
Founders
Assess whether NotebookLM's source-grounded model offers a template for building differentiated AI products in your space.
NotebookLM demonstrates that grounding AI in user-supplied sources creates a distinct value proposition versus general chatbots.
PMs
Map how the addition of visible reasoning changes the user experience for both casual and power users, and identify whether similar transparency features belong on your roadmap.
The update signals that AI transparency is becoming a competitive feature, not just a safety checkbox.
Investors
Monitor Google's continued investment in NotebookLM as a signal of whether source-grounded AI tools represent a durable product category.
NotebookLM's near-absence from I/O 2026 followed by a major update creates ambiguity about Google's strategic commitment.
Operators
Pilot NotebookLM with a research or knowledge-management team to test whether source-grounded AI improves information retrieval and synthesis workflows.
Multiple reviewers report that NotebookLM excels at transforming user-supplied information, making it a practical tool for internal knowledge work.
How to test
- 1Go to notebooklm.google.com and create a new notebook.
- 2Upload 2-3 source documents into the notebook.
- 3Ask a question that requires synthesizing information across multiple sources.
- 4Expand the visible-reasoning panel to inspect the model's intermediate thinking steps.
- 5Generate an Audio Overview from the same sources and evaluate its accuracy and usefulness.
- 6Compare answer quality and trust with reasoning visible versus hidden.
Caveats
- NotebookLM grounds responses in uploaded sources, but hallucinations can still occur—verify critical claims against originals.
- The recent update may change familiar workflows for existing users.
- Audio Overviews can take time to generate and may simplify complex source material.