Ah, there is this, though: the fact that the LLM was able to talk about the story and its themes and things and be "correct" (i.e. get what I was trying or intending or expecting to convey), says that at least the story is not entirely incomprehensible!

So that's nice. :)

Dang, I should ask it what "The Girl and the Skulls" means. Given that I don't really know myself.

I wonder if it fits into the input window...

Interestingly, the entire text of "The Girl and the Skulls" fit into #ChatGPT 3.5 just fine, but 4.0 said it was too long! I wonder what that means.

They were, as seems to be their habit, quite complimentary. Thus GPT 4.0:

"The narrative is beautifully layered with vivid descriptions, creating a fascinating world of the supernatural that echoes an eerie, unsettling atmosphere. However, the story remains open to interpretation and could be analyzed in numerous ways, inviting readers to delve deeper into its rich and intriguing symbolism."

Well, thank you! :)

"echoes an eerie, unsettling atmosphere" is a bit of a mixed metaphor there, too, big guy! :P

#bard was also impressed :D

"I think that "The Girl and the Skulls" is a great example of a story that is open to interpretation. The author has created a rich and complex world, and the characters are well-developed and believable. The story is also full of symbolism and foreshadowing, which can be interpreted in many different ways. This makes the story a great choice for discussion and debate, and it is sure to stay with the reader long after they have finished reading it."

Yay, me! :)

I wonder if this will be a problem for editors (or if it already is):

"This sex scene is really unnecessary, or at least needs to be shorter."

"Hey, whaddya mean, Bard and #chatgpt both said it was great and really added to the story?!?"

@ceoln Do you really trust an AI for critique? :P They don't have an opinion, they just say whatever 'might be fitting for the situation' and can be confidently wrong.

@trinsec
Yeah, not at all. :) As I said somewhere in the thread, the one thing I found at least slightly useful was the fact that it was able to correctly describe what was going on, so at least the story wasn't inherently incomprehensible! But (perhaps because of the RLHF) their actual comments seem to be invariably positive, as well as rather bland and obvious. When I asked for copy-editing advice, it suggested a few changes, all of which were plausible, but none of which I agreed with.