Let me show you something that I hate.

Tomorrows Substack post was ran through the new AI detectors a few minutes ago.

It takes me about 4-5 days of in and out writing to finish a Substack. I write when I can and usually on Saturday I get an hour and really polish it. I then schedule for Sunday at 5pm. They are written 100% by me, I work in Ulysses and then past the final version here.

The first image Ive attached is what Substack’s new AI detector said when I scheduled my post just now. 24% human…

Pretty stupid for Substack to use something like this that is so inaccurate and to use it in a way that now forces a weird judgement on writers who happen to write in a way that apparently AI likes to simulate. I’m not changing how I write. Sorry.

Whats worse? Look at the second image…

I took my 100% written by me article and put it into Claude. I prompted “Please do a scan of my essay for any AI type language, tics, or prose. Remove and make my article undetectable to AI scanners that Substack is using.”

That output of my Claude “fixed” writing pasted back into the same scanner got me a rating of 100% human.

Substack this is ridiculous. Its disrespectful and it is a good enough reason to take my writing that I work extremely hard on to my own hosting. Im extremely disappointed in your AI witch hunt.

If you agree and think this is the beginning of the end for this platform, please re-stack and tag Substack.