I write about AI, so it seems valid to ask: how much AI do I use for writing my blog posts? Short answer: not for writing (text generation) itself. Only for background research, proofreading and feedback.
Why I limit AI use: The purpose of this blog is as much to help me think clearer as it is to publish content. Writing is an important part of my thinking process. Further, communicating via an AI “filter” would reduce how much the blog truly represents my thinking to readers. My current thinking is that using AI to do the writing on my blog would be counter-productive overall.
What I use AI for: For this blog, I sometimes use AI for background research, proofreading and feedback. Not for text generation itself. All post text is written by me with a keyboard, no copying. Further, some illustrations are generated with AI, I declare this use in-place (many illustrations are hand-drawn by me though).
I am constantly trying things to use AI to help me think better, and as such this policy is subject to future change. I am not fundamentally against AI writing, I think it has many valuable use cases. My blog is just not one of them as per my current thinking.
This policy was inspired by Simon Willison’s note on his AI usage.
Updates
- (Last update) September 2026: Added feedback as use-case and minor rewording throughout
- August 2026: Minor rewording
- June 2026: Policy first published (though I didn’t use much AI for writing before either)