Google hasn't loudly announced a "12 Days of Christmas" event, but instead is steadily delivering every product update for the Gemini 2.0 era: the updates are substantial and take effect immediately.
This time, the update is for their surprise star product: NotebookLM.
First, there is a brand-new look:
The chat area has been enlarged this time, making it much clearer compared to the previous layout.
Secondly, the Audio Overview feature has been enhanced. Now, audio podcast sessions can exceed 60 minutes.
At the same time, a "Join" interactive feature has been added to Audio Overview, allowing users to directly interrupt the current conversation to ask questions, with the "AI Host" providing instant answers. The effect is truly stunning.
In fact, ever since OpenAI and Google integrated advanced voice capabilities into their models, I have harbored deep doubts about the future value of AI education apps similar to "certain fruit-named" ones. After seeing this Audio Overview feature in NotebookLM today, those doubts have deepened further.
Until now, I have maintained that AI applications should be "All-in-One" experiences based directly on model or hardware capabilities. Before the likes of Apple, Google, and OpenAI, other companies had little chance. However, as some friends recently reminded me, I began to reflect on whether this line of thinking and usage was too niche or too far ahead of its time.
Yet, seeing more and more developers from Google and OpenAI release application forms almost identical to what I envisioned, I am more convinced of my judgment: the mobile app era was about business model innovation—ultimately, whoever solved a pain point commercially could win massive market opportunities. In the AI era, capabilities are almost entirely based on the model itself and the existing ecosystem. So-called "innovations" built on top of big-company models actually lack foundation and vitality.
Opportunities are increasingly consolidating among a few key players.
Back to other updates in NotebookLM: processing speed and model capability.
With the launch of the Gemini 2.0 Flash model, NotebookLM's response speed has improved significantly, and its reasoning ability has also seen a notable boost.
More response adjustment options have also been added.
Additionally, a "Chat Only" sharing mode and analytical functions have been introduced.
The only thing I'm still a bit confused about is that many of the features I can already use are supposedly part of the "Notebook Plus" tier.
Based on my account type and Google's descriptions, the NotebookLM I’m using should be the Plus version, though I haven't seen any explicit confirmation or prompt.
Regardless, it shows that Google is providing significant value.
As I've said before, Google might have already claimed dominance over the AI application market for 2025.