How I Use (and Don't Use) AI
Everything published here begins with me. The ideas, arguments, and first draft come from my own experience and the work I have done, including the occasional rough edge that proves a person was here.
Writing is how I work through an idea until I understand it. An LLM can produce competent prose, but it cannot contribute my point of view or replace that process of thinking. If a post here sounds worth reading, I want that to be because it says something I genuinely mean, not because a model made it sound polished.
I do not treat AI as forbidden. Once I have a draft, I sometimes use it as a light copy-editing tool: to spot an awkward sentence, test whether a point is clear, or find a better word. The result should still feel like that greninja I raised myself: shaped through the actual grind, quirks and all, rather than traded in fully levelled.
What I Use AI For
- Checking grammar, spelling, and phrasing
- Suggesting alternative wording. As a non-native English speaker, I value the extra vocabulary without letting it flatten my voice
- Sanity-checking whether an article's argument follows clearly from one point to the next
- Brainstorming titles, teasers, tags, and small improvements to the site, which I rework to fit my own voice
What I Do Not Use AI For
- Generating the original ideas, perspectives, or conclusions in a post
- Writing a first draft or complete paragraphs and sections for publication
References
These AI disclosures helped shape how I think about this policy: