Writing This Blog With AI Help, Grounded in My Own Experience

I’m fairly open about this: a lot of posts on this blog start as a brainstorm with AI. But there’s a line I keep strict between “AI-assisted” and “AI-written.”

What I actually do: I give the AI real context — technical decisions I actually made, problems I actually ran into, code that actually exists in this repo. AI helps me structure a first draft faster. But every fact, number, and conclusion in it gets checked against the actual code or my actual experience before it’s published — none of it is AI invention that sounds plausible but never happened to me.

Why this matters to me: if content is purely AI-generated without editorial oversight, two bad things happen. First, the quality gets generic — it could be anyone’s writing, because it’s missing the specific details only I would know. Second, this is also relevant to something concrete like AdSense approval — Google explicitly states that AI content without human oversight typically fails review, because it isn’t genuinely “high-quality, original, and attracts an audience.”

How I keep my own voice: I always rewrite the opening and closing manually, because those are the parts that show how I actually think, not just how information gets organized. Specific details — project names, numbers, why I chose A over B — get filled in by hand, because that’s what makes a post impossible for anyone but me to regenerate.

AI is a tool that speeds up the process, not a replacement for experience. As long as I stay in control of what’s actually true and worth sharing, the result stays authentic.