Last year, Google introduced NotebookLM, an experimental AI notebook that blends LLMs and user notes to help you comprehend a topic by summarising, answering questions, and more.
NotebookLM employed text to summarise and explain source materials since its inception, but it can now do so audibly.
The function is intended towards persons who learn best by listening to explanations rather than reading them.
On Wednesday, Google revealed that NotebookLM, their AI note-taking and research tool, will now include an “Audio Overview” option.
Audio Overview provides users with additional means to consume and interpret the material in the papers they have submitted to the program, such as course readings or legal briefs.
In a blog post introducing the launch, Google provided an example of the AI host addressing the contents of a previous NotebookLM Keyword article.
AI-generated virtual hosts will use conversational speech patterns to provide you with a summary of the information you have supplied.
The hosts will discuss facts or fascinating subjects from the source material, and they may employ metaphors to clarify complex concepts.
According to Google, listening to these chats can help users discover new connections between their works or obtain inspiration for their drafts.
The voices are remarkably lifelike, and it sounds like you’re listening to a podcast with two hosts discussing a topic.
To get started, go to the NotebookLM website or Google Labs and select the ‘Try NotebookLM’ button.
Signing in requires a Google account, but the experience is free.
Then, to generate an Audio Overview, open an existing notebook, click on the Notebook guide, and select ‘Generate’.
An audio discussion can take up to five minutes to construct, and you can currently only generate chats in English.
You may also download the talk and listen to it on the move to acquire a better understanding of your source materials.
We earlier reported that Apple has opted to use Google’s AI chips, known as Tensor Processing Units (TPUs), to train its artificial intelligence models, bypassing Nvidia’s highly sought-after GPUs.
