Jeff,

Thanks for the response. :-D

If you're referring to the distribution of the 16 pictures for Mr. Midjourney (I'm assuming, since it's the only place I use the word "fluke" in the article), then we're both on the same page; it's not actually a fluke.

(I try to do my due diligence, especially where statistics are concerned; if it was a fluke, this article would have been very different).

But one of the challenges of walking people through an article like this is figuring out how to present results that are both illustrative and definitive. My way of doing that, here, was to first give people a clear visual using N=16 pictures, then acknowledge that it was a small number (for the few readers with enough grounding in statistics to wonder if the outcome was due to the small sample size), and then provide a more definitive tally using N=100. To do that, I asked the "fluke" question purely for the rhetorical purpose of leading into the next piece of evidence.

If I had tried to use chi-square tests and p-values and such I probably would have scared off readers. :-D