I Keep Teaching AI Who I Am and What I Want
I stopped throwing away my corrections. Now they become skills built from my own preferences.
Same pattern, every time, no matter what I’m working on with Claude — a pitch deck, a blog post, a piece of code:
- I give it input.
- It gives me something back.
- It’s not quite right.
- I adjust. Then adjust again. Then again.
It’s not that the output is bad. It’s that it’s generic — competent, reasonable, and not mine. Getting from there to something I’d actually ship takes round after round of correction. That loop used to just feel like friction. Annoying, a little bit crazy-making, the tax I pay for using the tool at all.
What I actually discovered
The corrections aren’t waste. I never sat down and wrote out what I actually want — it only ever showed up in the moment I rejected something.
The correction log is the most honest description of my own taste that exists anywhere.
Every one of those back-and-forth conversations is a complete, timestamped log of:
- what Claude produced by default
- exactly where I pushed back
- what I replaced it with instead
That log has been sitting there the whole time. I was throwing it away the second I got the deck, the post, the code I wanted.
The move
So instead of just closing the chat once the output is good, I now ask Claude to look back at its own transcript:
- Review this conversation.
- Find every place I corrected you.
- Pull out the pattern — not just what changed, but what I kept changing.
- Turn that into a skill: my structure, my wording, my do’s and don’ts.
- Load it automatically next time I do this kind of task.
Review our conversation. Find the places where I corrected your output, and pull out the pattern in what I kept changing. Turn that into a reusable skill — my structure, my phrasing, my do’s and don’ts — that you can load next time.
The insight isn’t the tool or the deck or the post. It’s this: the pattern only becomes visible after the work is done, never before it.
What it’s not
- Not a perfect skill file, built once and finished.
- Not a plan I made in advance of what I’d need.
- Not a fix — the next output still needs edits.
What it is
- A small, honest upgrade to my starting point.
- One correction cycle mined and folded back in, so I’m not re-litigating the same shape every time.
- Proof that watching what I actually did beats guessing what I might need.
I tried it once, on the pitch-deck pattern. The next draft was better on the first pass — not done, just closer. I still edited it. But I started from my structure instead of a blank paragraph, which meant I spent the time arguing about the actual content instead of re-fighting the shape again.
Baby steps. But now every messy correction cycle is a source, not just a cost.