The software generates a podcast referred to as deep divingwhich includes a female and male voice discussing what you have uploaded. The voices are astonishingly practical: the episodes are plagued by little human-sounding phrases like “Man,” “Wow,” “Oh, proper,” and “Wait, let me get this proper.” The “hosts” even interrupt one another.
To check it, I copied every story from MIT Expertise Overview‘s a hundred and twenty fifth anniversary quantity in NotebookLM and had the system generate a 10-minute podcast with the outcomes. The system selected a few tales to concentrate on and the AI hosts did an important job of conveying the overall high-level essence of what the difficulty was about. Pay attention.
MIT Expertise Overview a hundred and twenty fifth Anniversary Version
The bogus intelligence system is designed to create “magic in trade for somewhat content material,” stated Raiza Martin, product chief at NotebookLM. saying in X. The voice mannequin is meant to create emotional and interesting audio, delivered in an “upbeat, hyper-interested tone,” Martin stated.
LM pocket book, which was initially marketed as a examine softwarehas taken on a lifetime of its personal amongst customers. The corporate is now working so as to add extra customization choices, similar to altering the length, format, voices and languages, Martin stated. It’s at present purported to generate podcasts solely in English, however some customers on Reddit managed to get the software to create audio in French and Hungarian.
Sure, it is cool, virtually scrumptious even, but it surely’s additionally not resistant to the issues that plague generative AI, similar to hallucinations and biases.
These are a few of the fundamental methods individuals are utilizing NotebookLM to this point.
On-demand podcast
Andrej Karpathy, member of the founding workforce at OpenAI and beforehand head of AI at Tesla, stated in X that deep diving It is now your favourite podcast. Karpathy created his personal AI podcast sequence referred to as Thriller taleswhose aim is “to uncover historical past’s most intriguing mysteries.” He says he researched matters utilizing ChatGPT, Claude, and Google, and used a Wikipedia hyperlink for every subject as supply materials in NotebookLM to generate audio. He then used NotebookLM to generate the episode descriptions. It took him two hours to create the whole podcast sequence, he says.
