What I'm changing my mind about.

About a month ago I wrote a post called What I was wrong about. It was about the personal use case for AI being much bigger than I’d thought it would be when I was working in tech and listening to the loud version of the conversation.

That post got more replies than almost any other I’ve written.

A lot of you asked me, in different ways, what else I thought I might be wrong about. I said I’d come back when I knew.

Today I want to flag three things I’m changing my mind about, two months into writing this newsletter. None of these are settled. All of these are mid-reconsidering. But I want to mark the calendar on them, because part of what I’m trying to do here is be a tech person who keeps changing his mind in public, instead of one who keeps pretending he was right all along.

Six weeks ago, when I was planning this newsletter, I told myself I wasn’t going to do tutorials. Tutorials were what every other AI newsletter on earth was doing. I wanted to be different. I wanted to find the voice of this thing first, in stories and observations, and earn the right to tell anyone how to do anything.

That instinct was right. The first month being voice-only was correct, and I’d do it the same way again.

But two weeks ago I published A bedtime story you can make tonight, which was a full tutorial. It got more replies than any post I’ve written. Almost all of them said some version of: thank you, I needed someone to walk me through it. A few of them sent me photos of the printed books they made.

I am now reconsidering whether tutorials are the productivity-bro thing I had filed them as. I think tutorials done a particular way — small, specific, with a real person on the receiving end — might just be a different way of doing what we’ve been doing all along, which is helping someone make something for someone they love. I’m going to keep writing them. Probably one every two weeks. Mixed in with the stories.

When I started this, I had a picture in my head of who was reading. The picture was: curious adults, thirty-five to fifty-five, urban, not in tech, slightly intimidated by AI, slightly skeptical of where it was going.

That picture was wrong in two specific ways.

The first wrong thing: a much bigger chunk of you are in tech. Founders, engineers, designers, product people who keep telling me, I sit in the loud version of this conversation all day, and your newsletter is the only place I can read about what it actually feels like to use this technology with the people I love. I did not expect that. I thought tech people would find AI for Fun too soft. Many of you have written to say it’s the opposite — that it’s the only place where the technology feels like it’s about anything other than work.

The second wrong thing: a much bigger chunk of you are much older than I thought. I have readers in their seventies. Several. They’re using AI to write to their grandchildren, to translate poems from their first language, to record their own life stories. I had filed older people as the AI conversation’s blind spot. They’re actually one of its largest hidden audiences. I’m going to write more for them, more carefully.

When I started, I thought this was a project of mine — a place to write down the version of AI I cared about, in the hope that the right readers would find it and we’d grow a community around it slowly.

I was thinking of myself as the center.

Two months in, I don’t think that’s right. What’s actually happening, in the replies, is that you are using me as the excuse to do the thing you’ve been wanting to do. You’re making cards for your dad. You’re writing letters to your aunt. You’re recording your grandmother’s stories. You’re using each post here as a small permission slip to do one more thing for one more person you love.

That changes how I should write going forward. I’m not the protagonist of this thing. You are. I should keep telling stories — they’re how I see what’s happening — but I should remember the stories are not the work. The work is happening on your end. In your kitchens. To your dads.

I have a piece I’m writing about a song my cousin made for his eighty-year-old father in his hospital room, and I want to tell you about it. It’s a hard one. I’m still figuring out how to write it.

In the meantime — make something for someone this weekend. Small. Specific. Just one thing. That’s where the actual newsletter lives, on your side.

I’m really glad you’re here.

— Tapan

PS — What are you changing your mind about right now, about AI? I’m collecting these. The honest ones are the most useful. Reply and tell me.