I recently tried NotebookLM by Google.
How it works?
You upload a set of files into it and can extract insight from that content by asking questions.
More or less, it works like chatGPT, except you set the limits of the source from which your question is to be answered. That means it will not respond from a general point of view but from the content of the source you have given to it.
When it responds, it comes up with a summary or insight with the citation to the source.
It is most suitable for learning, brainstorming, and doing research surveys.
Audio notes.
One good feature is that you can produce audio notes similar to a podcast. This feature is called Deep Dive. I have not used it, so I have no idea how good it is.
Comparison.
There’s no appropriate way to compare this with chatGPT. As chatGPT responds to each question from the uploaded content along with its pre-trained knowledge.
But NotebookLM utilizes data only from files you feed it. The file could be a pdf, image, or a set of files in a different format.
I am just getting my hands on this new tool.
One area where NotebookLM shines better than other AI is its ability to cite the source. I have seen a similar feature in SciSpace. In ChatGPT, it cites from everywhere.
NotebookLM supports multiple citations, which means it can cite multiple sources that discuss the same idea. Its contextual understanding is better than that of other AIs I tried.
I suggest you test on your own. I do not have any data or facts to judge the accuracy of this tool’s output.