Hey y’all! A few weeks ago, I came across an open-source repo — Blader’s Humanizer (14k+ stars). Around the same time, one of my clients was being quite strict about AI detection scores (GPTZero, specifically). Out of curiosity, I ran one of my blog posts through the humanizer. My score dropped from ~30% to 9%. I was honestly surprised. It felt like months of frustration just… melted away for a moment. But then I read the output. And it didn’t quite sound like me. That stuck with me. So I decided to properly stress-test it. I took 10 AI-generated blog posts and evaluated the humanizer across 9 editorial axes, scoring everything myself. Took me a little over a week. I ended up writing a detailed breakdown on Substack. Would really appreciate your support and especially any constructive feedback 🙏