Our first blog post got called out for AI-isms. Touche.

Published our first blog post this week on why AI writing all sounds the same. It got shared on Reddit and some of the feedback was honest. A couple of members pointed out that the article itself had AI patterns in it. They were right. Triple parallel structures, hedge qualifiers, the same "not X but Y" reframe used three times. We went back, found the tells, and fixed them.

We use AI to write content for Noren, these patterns were possible because the brand Noren is still building a voice. In a couple of months of writing and rewriting, we will be able to use Noren, to extract how Noren writes (brain freeze).

The lesson: AI-isms aren't bad because they're AI patterns, they're bad when they replace yours. The repetition is what makes it detectable, same structures, same rhythms, same transitions. Every writer repeats patterns too, that's what makes a voice. Go read Paul Graham's blog from 5 years+ ago, he writes like today's ai. The real issue is that AI repeats the same patterns for everyone, which frustrates readers.

We're building Noren to fix exactly this, extract a writer's actual patterns and feed them back into the generation so the output sounds like them, not like default GPT.

First week, first blog, first public feedback loop and first updates to Noren voice. More posts coming.

Blog: https://usenoren.ai/blog/why-ai-writing-sounds-the-same