I use NotebookLM as an attack layer before I publish anything.
Not for summaries. Not for research. I upload my draft, let the audio overview run, and listen to two AI hosts interrogate my argument. If the narration exposes a gap — a claim that doesn't land, a section that wanders — I go back and fix it before the article goes live.
It works. Better than re-reading. Better than asking a colleague. The distance of hearing your argument spoken back at you by something that doesn't share your assumptions is genuinely useful.
One thing is missing.
The voice. My voice.
