The device creates a podcast called “, featuring a dialogue between two voices – one masculine and the other feminine – that discusses the topic or content you had uploaded. The voiceovers are astonishingly lifelike – peppered with colloquialisms such as “Man”, “Wow”, “Oh proper” and “Hold up, let me get this right.” The “hosts” even occasionally interrupt each other.
To verify the results, I replicated each narrative in NotebookLM and had the system create a 10-minute podcast based on the outputs. The selected narratives were honed in on, with AI facilitators delivering an impressive summary of the key issue’s essence.
Raiza Martin, NotebookLM’s product lead, describes their AI system as capable of generating “magic” from minimal input, on X. The voice mannequin is designed to craft captivating audio, delivering a lively, hyper-engaged tone that evokes strong emotions, according to Martin.
The Notebook brand has built a lifetime of loyalty with its discerning clientele. The corporation is now investing in expanding its customization options, including adjustments to font sizes, formats, voiceovers, and language support, according to Martin. It is rumored that the device will initially produce podcasts exclusively in English. Nevertheless, several adventurous users on Reddit have reportedly successfully coaxed the device into generating audio content in other formats.
While generative AI models like this one may exhibit a certain charm, they are not entirely impervious to the pitfalls that beset their kind, including hallucinations and biases.
Notable applications of NotebookLM thus far include:
Andrej Karpathy, a founding member of OpenAI and former director of artificial intelligence at Tesla, has recently revealed that he’s become enamored with his own favorite podcast. Andrew Karpathy’s own AI podcast archive, titled ?, aims to “revel in the fascinating enigmas of history.” He reveals having leveraged ChatGPT, Claude, and Google for research, while relying on Wikipedia links from each topic as source material for NotebookLM to produce audio. Utilizing NotebookLM, he crafted compelling episode summaries. Within a span of just two hours, he claims to have curated an extensive podcast collection.
