How I use AI

(This is an /ai page inspired by Derek Sivers. It goes back to the manifesto by Damola. If you have your own site, you should make one, too.)

As of October 18th, 2026, this is how I use generative AI/LLM tools in my work.

In short: I make everything here myself, with AI as my assistant. I start it, and I finish it.

Writing

I think out loud. I dictate stream-of-consciousness notes, often on a walk, often messy, sometimes exactly the way I would write. AI transcribes them and cleans them up the way a good editor would. Then I read and edit every line. The ideas, the stories, and the final word are mine.

I also use it like a smart friend at the next table: to research a topic, summarize something long, or find the thread in my thinking.

What I never do: “Write an article about X.”

Programming

Here I hand over more. AI writes parts of our apps, tests my code, and catches what I miss. I make sure I understand what it builds, and I am very opinionated about the path on how we get there. Some days it amazes me; some days it frustrates me to no end, in roughly equal measure.

For writing, the words are the work. For software, the work is a thing that runs: a website that loads, a member who can watch a class on their phone.

Photography

My Leica or Canon cameras go almost everywhere with me. I document; I don’t generate.

My photographs. I keep my work as true to the moment as I can. I use masks, small cleanups, and the same spot fixes I’ve always made with the brush, clone, and heal tools in Lightroom and Photoshop. Some of those tools use AI now, and occasionally I use content-aware fill the same way. I don’t remove whole buildings, recreate a scene, or replace a sky or background with AI. If I call it my photograph, it’s what was in front of my lens.

Images that just need to fit. Things like a website header or a banner are a different job. Occasionally I extend a landscape to fit the layout or tidy a corner so the text can sit on top. That’s functional design. I’d never enter these in a competition or hang them on a wall.

Videography

I have humans in front of the camera (including myself). I enjoy cameras far too much to let a machine replace them. AI helps with transcripts, subtitles, and switching between cameras in interviews. I watch everything, cut the pauses myself, and read every subtitle, because AI gets plenty wrong.

Translation & Voice

This is the one I’m unsure about. AI dubbing could bring our work to people in languages we can’t yet afford human translators for. That’s a real chance to help more people. It’s also a real question. I haven’t decided yet.

Where I stand

I like being early to new tools. I played with Photoshop in the early 2000s, and I shoot digital, not film.

I’m also conflicted about the ethics, about what counts as human work, and about what we lose when it gets too easy. So I use AI deliberately, in some places and not in others. My rule is simple: I hold the pen.

This will probably change. When it does, I’ll update this page at this permanent URL: ChrisSpiegl.com/ai

If you see it differently, email me. I’d like to hear it.