E6: He Looked Like You to Me; S15: What Did He Look Like to You?

I have a slight fear that explaining this episode will cause me to add something else to it, and I don’t want to. It already seems finished.So this is just information about how it happened.About two weeks ago, I spent an afternoon being a gremlin on Substack. I fed some of what I was doing into AI, asked it to find patterns, and eventually had it make a song out of the whole thing. Then I asked it to criticize that song. Then I criticized the criticism. Somewhere in there, an old song I wrote around 1988 called “Fate’s Bait” appeared to belong with the others.I put the four songs in a playlist called “Gremlin.”Then I forgot about them.Today I opened the playlist and barely remembered what was in it. I listened to the songs once while walking home. For much of that first listen, I could hardly hear them because I was walking beside traffic and I don’t wear in-ear headphones because they make my ears itch.I listened to the first song one more time.Then I started talking.Everything before the songs in this episode was improvised while I walked. There was no outline, no notes and no plan for where I was going. I knew vaguely how the songs had been made. I remembered pieces of my own history as they occurred to me. I talked until the thing reached an endpoint that felt completely obvious, and then I stopped.What interests me now is not whether that counts as impressive improvisation. I’m more interested in what a person is doing when something like this happens at all.Maybe preparation is the wrong model.If you have spent decades repeatedly noticing certain things—how categories work, how language changes what it describes, how individual experience looks different when scale changes, how memory ignores the lines we draw through our lives—perhaps you do not need to retrieve those ideas deliberately. They are no longer a collection of ideas. They have become something closer to the machinery doing the retrieving.So when I speak without knowing what I am going to say, perhaps I am not inventing a structure in real time so much as exposing one.The particularly strange part is that, afterward, when I gave the transcript and songs to AI, it pointed out that the improvised monologue had already introduced, through autobiography, many of the same problems the songs would later make explicit: categories, framing, scale, the fantasy of a closed chapter, the problem of an authentic self, observation as evidence, and the persistence of the person on the other side of a line.Apparently I had heard all of that while barely hearing the songs in traffic.Or perhaps I didn’t have to hear it.Maybe I concoct things out of myself.And whatever comes out will tend to resemble me.And, not out of a responsibility for transparency, I'll note that the above was finalized by AI. If you listen to this episode, you'll experience how I talk.I talk in the same way to AI, but with even more-flowery language. I groove languidly like Billie singing in a bar. Or at least that's how I feel.Maybe I mean Billy Tipton.