Now that Fable is back on the market, I did an apples-to-apples test with it.

While I purposely avoided using any AI to write Stolguard, I did feed the finished manuscript into ChatGPT and Claude prior to publication and asked them to analyze the characters, plot, themes, prose, genre, and overall thoughts. Like a robot beta reader among my myriad human beta readers. They did a half-decent job, like a smart high-schooler who read it but skimmed certain parts, made a few objective errors (including a big one involving all the nuance around Jade and what she’s up to and why, across the two interwoven timelines, iykyk), but otherwise got the gist of the book with some random inconsistencies.

I did that with Fable now. Just gave it the text and asked it to analyze it along those same areas. Its overall analysis conclusions were quite similar, but it avoided making those objective mix-ups around the most complex parts. It understood the interwoven timelines fully. Only with some follow-up questions did it start to slip a bit with some clear mistaking of one character for another or forgetting about certain scenes as its context window failed to keep up.

Definitely a step up in accuracy, based on a text I know well.