I have used Google Notebook LLM to create a podcast based on two of my notes on Broadcom. This was 25 minutes Podcast was fine for the first 23 minutes. After that the Podcast started talking about a third document (which is unknown to me) and started talking complete nonsense. The underlying model was hallucinating.
This is an important lesson about the limits of AI and how its results have to be checked very carefully. In this case, please do not listen to the last 2-3 minutes or at least do not take it seriously as it is wrong and misleading .
Nevertheless, these podcasts are potentially a good convenient introduction to a topic. They are not a substitute for reading the written note, as the latter has charts and other important data which cannot be conveyed in a podcast.
Share this post