The easiest way to tell that a SaaS blog used an LLM is not an AI detector. It is the moment the prose starts sounding like every other SaaS blog.
The Turing Window
I made this term up because I wanted a name for the awkward period when a new AI writing habit feels novel, then suddenly becomes an obvious tell. A Turing Window The period between the introduction of a novel AI-generated linguistic or behavioral pattern and its widespread recognition as an indicator of AI authorship. .
The reference is to the Turing test, but the direction is reversed. The question is no longer whether a machine can sound human. It is whether a human can avoid sounding like the machine everyone else is using.
Every AI-ism has a half-life
Right now, the tells are easy to spot:
- “It’s not X. It’s Y.”
- an em dash every other sentence
- “the X situation”
- a suspiciously tidy list of three adjectives
None of these is forbidden. The problem is accumulation. Once a pattern becomes common enough, it stops reading as style and starts reading as evidence.
Every AI-ism will age badly. Not because the sentence becomes ungrammatical, but because the pattern becomes a timestamp. Readers will know which generation of writing tool was popular when you published it. The sentence stops sounding polished and starts sounding like a fossil.
Use AI. Edit like a person.
This is not an argument against using AI to write. Use it for 99% of the first draft if that gets you moving. Give it your notes, examples, and ugly half-sentences. Let it make something you can react to.
Then do the important part yourself: delete the throat-clearing, replace generic examples with exact details, break the perfect list, and remove the em dashes you did not consciously choose.
Read it out loud. You will hear the phrases you would never say.
Use the machine to get to a draft. Then edit until it could survive outside the Turing Window.
