This evening I played around with NotebookLM for the first time.
A feature I noticed straight away, was the “Audio Overview”.
My understanding is, for any chat thread with NotebookLM, you can generate an audio overview, notably, in the form of two people discussing it, podcast style.
I decided to throw one of my recent posts, “One Mind, Many Hands” to it and see what would happen.
I gave it no additional context or prompting, this was a pure “one-shot”.
The AI itself is impressive. It feels and sounds like a podcast. It definitely has that NYT Daily style cadence and tone.
There are weird quirks, to be sure. It’s not perfect.
And I don’t know if I’d listen to this.
Or would I?
Nevertheless, it’s extremely interesting, especially if this gets better, which you can only assume it will.
If it becomes more able to be creative, to be interesting, to go on tangents and to argue, then I could see this being interesting.
As always, the worst-case-scenario is just an endless slop machine.
But as Scott Belsky says, “AI will both lower the floor and raise the ceiling for creativity”.
I’m very certain this sort of text-to-audio technology is going to lead to a ton of interesting use cases and new formats.
