Makes sense! Maybe my confusion (which was more about the format than the topic), could be solved just by some subtle sound effect to mark the switch between you talking to the guest and you talking to the listener? Since there's no video in a podcast to give this context, the listener has to exert some effort trying to deduce it from other clues.
For example, Accidental Tech Podcast inserts a small jingle at the start of a sponsor break whereas Talk Python inserts a sudden awkward silence in the conversation. It took me many episodes of the latter to stop wondering if the player had crashed, or if the interviewer lost his train of thought, or the guest was offended, etc. The jingle lets the listener know what's happening, which gives a more relaxing listing experience.
So some easy way to distinguish the interview from the narration would be useful, I think.
But maybe I'll be less confused once I've listened to a few more episodes. A new podcast always takes some getting used to. Thanks for making it!