Aadil Mamujee

I’ve been writing on my personal blog since 2012 (122 posts over 14 years) and many of them are decent, but they could have been written by anyone.

I want to be a better writer. I’ve read books, hired an editor, and asked friends for honest feedback. But it’s hard to be objective about your own work, and harder still to know if you’re improving. I’m not optimizing for distribution or shareability so views/shares are not the right metric. I care about the quality of thinking, documenting the things I’m passionate about, and sounding like myself when I do it.

The posts I think are strongest – the neuromuscular condition one, a virtual funeral for my great uncle Taher, turning forty – are the ones where I wasn’t trying to sound like a writer. The weakest ones usually started with a framework from a book I’d just read or were structured like product memos.

I exported all my blog posts into a Claude project as context and got it to tell me what I did well, what my persistent issues were, and where my writing was strongest. I packaged the findings into a custom skill, which is a set of instructions Claude loads whenever I start a new draft. Now there’s a partner in my publishing workflow that has read everything I’ve ever published and compares the new writing to the old.

I dictate rough drafts into Superwhisper, work through structure with Claude, then run the draft through multiple models for feedback. The feedback is usually the same: go deeper into the things only I would know.

The obvious conclusion is that the writing quality improved and this writing system works. I’m not convinced. There’s a real risk that what feels like coaching is the AI quietly pulling me toward what all the AIs think good writing looks like. Reversion to the mean, dressed up as feedback. Sentences get shorter. Hedges get cleaned up, but so do the sentences that are awkward on purpose and sound like something I would say.

My wife read a few heavily AI-edited drafts and told me they sounded like an LLM wrote them, not me. Since then I’ve been second-guessing almost every edit. I should be the writer and the editor, not the model. But the model might also be right, and I don’t always know when I’m losing my voice versus making the writing better.

A PRD at work does not bother me in the same way. If Claude makes a product memo more structured and less like me, that’s a feature. But writing on my personal blog is different. The version of me that shows up in a post about my family or Japanese knives is the whole point. Smoothing it out with AI means losing my authentic tone and I often can’t tell if that is happening.

Take this post. The first version opened with biographical stats: how long I’d been writing, the metric I wasn’t optimizing for. The writing system flagged it as memo-shaped, and it was right. I rewrote the opening. It flagged a paragraph in the middle that had drifted into product-doc voice, phrases like “iterate faster” and “the feedback loop has collapsed from months to minutes,” and that came out too. It also told me the best line in the draft was “reversion to the mean, dressed up as feedback.” I liked that line too, but models are also mirrors designed to make me feel good, not be a true coach.

I don’t have this figured out. The system is good at catching things that are genuinely bad: hedging, bullet points that should have been sentences, generic frameworks. It’s less good at knowing when a sentence is awkward on purpose, when a tangent is the point, when the worst-flowing paragraph in a post is the one I most need to keep. That’s where I have to push back on the edits – sometimes I will, and other times I’ll cave and regret it later. There will be posts that come out of this process a little too “clean”, and I’ll know which ones when I reread them in a year.

I’d rather find out a draft is mediocre while I’m still editing it than six months after I published it, and the writing system gets me there faster than anything else I’ve tried. The cost is that the draft I’m editing is already, slightly, not mine. The version that gets published is the version that goes through the smoothing, and that version is closer to what the LLMs think is good writing – reversion to the mean.

For now, I’m treating the writing system as a reader I trust some of the time, not a tone coach. My voice is the thing I bring and the writing system is there to tell me when I’ve drifted away. The next five posts will tell me whether this new writing system is useful or just a very polite critic that agrees with itself.