Yesterday, I had an odd idea: what would happen if I got AI to make a podcast episode reacting to something I wrote?
Not just summarizing or explaining it — but talking about it. Like two friends discussing last night’s TV episode. Or coworkers chatting about a Slack thread you posted.
Basically: me, listening in on two AI hosts talking about me, my ideas, my blog… behind my back. Kind of.
Talk about cringe! But also… kind of thrilling. And that’s why I did it.
The spark
The idea came from a YouTube video by The Verge where the host, Andru Marino, demoed Google’s NotebookLM — a tool that can generate a full podcast episode from source material.
It sounds human. Too human. It’s got “ums,” awkward banter, and genuine-sounding laughter. But what stuck with me wasn’t the tech. It was how Andru used it.
At one point, he uploaded his own job performance review and asked the AI to make a podcast about it:
“It’s hard to handle criticism… but maybe it’s easier in audio form.”
Spoiler: it wasn’t!
That was the spark. If AI could reflect on his work — why not mine?
Why?!
I write to feel better, not to feel better about myself. I also write to become a better communicator. To better understand how my thoughts land — not just in my own head, but in someone else’s.
And that’s what this AI podcast experiment is really about:
- Practicing what it feels like to hear feedback about myself
- Understanding how my ideas and writing are interpreted
- Improving the clarity of how I express what I mean
Most of the time, I ask ChatGPT to analyze my writing — to be blunt, to not flatter, to criticize with honesty. That helps me see what’s working and what isn’t. This podcast format takes it further. It adds voice. Tone. Personality. Distance.
It lets me hear how my work might be talked about — even if it’s just simulated.
Mic check...
To test things out, I uploaded my blog post on The Cringe — that restless, uncertain feeling after you share something.
(Perhaps writing about it encouraged me to do something deeply cringeworthy.)
I didn’t want the AI to summarize it. I wanted it to react to it. To break it down. To review it like it was a film or album.
Here’s what I prompted it with:
“The blog’s author, Q (he/him). In this episode, the hosts react to and break down one of Q’s blog posts. They share their thoughts, questions, and personal examples.”
The process was simple enough.
- Add blog post URL
- Paste in prompt
- Click generate
- Wait 90 seconds
- Brace for impact
Press play
The podcast was generated: a 7-minute, 20-second audio file of two AI hosts talking about me, my ideas, and… goodness knows what.
The weather was nice that day. I mustered up the courage to go out for a run (running sucks). And what better soundtrack for a run than a simulated podcast episode of two artificial voices casually dissecting your thoughts?
I downloaded the file. Put on my shoes. Stepped outside. Gritted my teeth (for more reasons than one). Pressed play. And started running.
I ended up running 4KM that afternoon.
Next episode
Later on that day, I went further:
I uploaded multiple blog posts about how I take notes — including The CODE System, My Way, plus posts on stacking ammo, not forgetting, and making use of wasted time.
Here’s the prompt I used this time:
The blog’s author is Q (he/him). The podcast is called “Deep Dive.” Hosts: Lee (male voice) and Sam (female voice). In this episode, the hosts explore Q’s personal note-taking approach. What’s interesting? What stood out? What do they wish Q expanded on? They should share examples from their own lives. Then, simplify the takeaway.
Once again, it worked. And once again… I cringed. But I also listened. Multiple times. Once while sitting at my desk. Once while folding laundry.
It was surreal — and strangely fascinating — to hear “people” talk about my ideas this way. Casual, conversational. Like I had tuned into some parallel universe where my thoughts were being passed around, interpreted, reflected on.
It didn’t just help me hear what others might take away — it helped me hear what I really meant.
What I’m learning
One phrase stood out in the generated podcast. This idea that the writing feels "lived-in". That’s not something I’d ever say about myself. But it is something I hoped would come across. Hearing it made me feel… seen. Even if the “person” seeing me wasn’t real.
And this wasn’t the first time I’d heard it. It’s something ChatGPT has surfaced before — describing my writing as grounded, experiential, real. The kind of clarity that only comes from actually doing the thing.
So when I heard that same sentiment echoed from the podcast hosts, it caught me off guard. But in a good way.
Because even if the podcast is fake, the feedback feels real.
Want to hear it?
(Oh my goodness!)
Here’s the podcast episode where two AI hosts talk about me and my blog post (below).
And no — I haven’t fully recovered from listening to it. But I think I’m going to keep doing this. Not because it’s comfortable. But because it helps me grow.