I made an AI song about my son. It had his name in it and references to things he liked. It took about a minute. He enjoyed it, and so did I.
About a week later, the novelty had worn off. I didn’t return to it.
My wife pointed out what I’d missed: making a song with him could have been the point. I’d made something about him and skipped an experience we could have had together.
That became the question behind this episode: is the journey the point?
1. The song
The output wasn’t the problem. I had asked for the wrong kind of value. An entertaining artifact and time spent making something together are different things. One doesn’t automatically give you the other.
2. The summaries
I used book summaries and called it efficient. But part of what I enjoy about reading is finding an aside, an unexpected reference, or a book I didn’t know to look for.
A summary can help me orient myself. It doesn’t give me the same experience as reading a book I want to spend time with. Now I try to be clearer about which of those jobs I’m asking it to do.
3. The analysis
In my real estate work, a polished retail site analysis covered the obvious categories. The question I cared about was what it had missed: was another shopping center being planned within the trade area?
Years of doing the work had taught me to ask that question. A convincing report doesn’t prove the right evidence was gathered, or even that the important question was asked. I still need to check what lies outside it, including with people who know the local situation.
4. The card
A birthday note or thank-you begins before the words. It begins with remembering what someone means to you.
If AI writes first, I can skip that remembering. The sequence I prefer is simple: I write the specific, possibly awkward version. Then AI can help me clarify it without inventing the feeling or the memory.
5. The switch
The larger mistake was treating AI as something to turn on or off for a whole activity. I find it more useful to think about where to point it.
In 2020, I built bookshelves for our home library. Six weeks became six months. Better planning might have helped, but cutting, sanding, painting, and learning were a large part of what made it worthwhile.
AI could have helped me get ready to do the work. I would still have wanted to do the work.
Turn that into one decision
The journey worksheet helps you separate the outcome you want, the experience you want to keep, and the friction you’d gladly remove. Try it on one thing you’ve already started handing over.
The companion episode, five AI uses I’m never giving back, explores the other side: what happens when a tool opens an experience that wasn’t realistically available to me before.