This needs to be a quick note because the month is about to end, I’m sick again, and I want to get back to the ending of this book. (I am rewriting it pretty extensively from my first attempt. It’s not the climax that I’m changing, it’s the epilogue. It’ll be a bit long, certainly longer than the one I wrote for Child of Fire, which irritated–in a good way, imo–so many people.)

There has been much talk about so-called AI and fiction lately. I say “so-called” because there are a lot of different kinds of software being labeled “AI”, and much of it (the maybe-possibly useful stuff) seems to be about analysis of scientific data. Not all of it is good, but people who know more than I do have managed to get worthwhile results (like fancy math proofs) from it. And that’s fine.

But when we talk about fiction, we’re almost always talking about large language models, and those are pretty clearly not good.

I don’t know if I’ve said this here before, but I have never used an LLM in my writing and I have no intention of doing so in the future. Never not ever.

First of all, the “AI” branding is fake. The text it spits out might seem to be sapient, but it’s not. The software has no intelligence. It can’t vet itself because it has no conscious awareness. It’s a plagiarism algorithm that is only good at guessing which word should follow next, based on the thousands of authors whose work has been stolen to create it.

Second, the last thing in the world that I want is a story, paragraph, or sentence that proceeds in the most statistically likely way. I’m not aiming for average. I’m trying to do something interesting and unique. Flooding Amazon with generic, machine-processed stories is only good for bilking people out of their hard-earned money. I work hard so that people feel glad that they bought my book.

Third, I sometimes see people bragging about how good they are with “prompts”, and I recognize them as the same guys who used to say they had a great idea for a book. Then they’d suggest I write it up so we could split the money fifty/fifty.

Ideas are cheap, and while I’m glad this particular kind of guy isn’t bothering authors any more, I’m sorry to see them dumping their pseudo-stories into the same marketplaces that actual human writers need.

Okay. Back to it.