Yog-Sothothery and AI

I've had fun doing the New Pantheon series of posts, although it's taken longer than I thought and I'm... ⅗s of the way through, I suppose. Two more posts to go to finish it. I've done a few things differently for those posts. I have a lot of AI generated images, which isn't new for me, and I generated most of them myself, but I did do some image searches of AI images that other people have generated and then uploaded somewhere. Not that I would feel bad about using images found online in any case, but I especially don't feel bad about using someone else's AI generated images, because... they're just AI generated images.

No, what is new is that I also used ChatGPT to generate drafts of some of the text. I didn't have detailed descriptions of any of the gods of the New Pantheon, but I knew more or less what I wanted them to be, so I threw a prompt in for each and asked it to whip up a little description, and for the most part, I liked my results well enough that it seemed easier to copy and paste and then do an editing pass (more for style than for content, but some of each) which I've never really done before. But I found it a useful shortcut. I suspect that as I did more of them, my editing got more radical, so that from a style perspective the later ones sound less like AI generated text. But, none of them are truly AI generated text, I used AI to generate a first draft, and because AI text usually reads like it was written by ... well, by AI, I had to edit those drafts anyway. I don't know that there's a reason to us this tool again in the near term, but I'm at least sufficiently happy with the results that I'll finish the next two New Pantheon posts that way, and keep the tool in my back pocket just in case I'm interested in using it for something else again. I'm curious if I use it to flesh out just a bit my rather high level and lacking in details 5x5 Fronts if the juice will be worth the squeeze. The important part of those is the high level outline; I don't expect the details to all be innovative and unique, so delegating some of that to ChatGPT is probably fine. We'll see what it does. 

I do get a bit annoyed with people who reflexively knee-jerk away from the word AI. Just because something is AI doesn't necessarily mean that it's useless slop. Just because something isn't AI doesn't mean that most of it isn't useless slop too. In general, my own personal experience is that AI mostly generates slop, but at least its slop of consistent quality, about the median in what is otherwise a sea of non-AI slop. And if you take AI as the baseline from which to improve upon, then you're already at least at the median of most of what is produced anyway. Not a bad place to start. Especially if you actually are capable of improving it and can use it to accelerate your output. And using AI to generate images for blog posts, YouTube videos and whatnot is now becoming pretty commonplace. AI is coming, no matter what you think of it. It's not going to live up to the hype of the boosters who claim it is a genius robot intelligence that can do everything better than you can, but it will 1) democratize certain tasks that are otherwise out of reach, like getting "good enough" art to make videos and blog posts more visually appealing (and eventually probably replace a lot of book covers and other applications, and 2) be a useful tool that can accelerate the productivity and output of people who find narrow opportunities to use it and can do so in a way that still gives them high quality output. I think it will really accelerate novel writing, for instance, turning details outlines into (admittedly, poor) first drafts. But that's still a major step and will cut a lot of time from people who are writing. I've also seen from Castalia Press, for example, that with a good editor, it can rapidly translate foreign works into English, for instance.

For instance, here's an AI rendering of my disco Sith Lord on Korriban from SWTOR, just for the heckuvit. 

I will also note that using AI this way is often kind of frustrating. I had to put in a few edits to the prompt and still didn't really get exactly what I wanted, but eh. It's close enough. It's a tool that gives interesting, yet not perfect results, obviously.

Anyway, I didn't mean to write a post about AI. I'm actually not sure what I intended to write a post about. Something vaguely Yog-Sothothery, I believe. So let's do that. 

I've been reading SPCM Sagas, i.e., adventure path or campaign-like products by Sandy Petersen's Cthulhu Mythos, a now defunct company that briefly made a bunch of 5e D&D products that brought Lovecraftian elements directly into D&D. Unlike other products he's worked on, such as Call of Cthulhu itself, which he created for Chaosium and which was published in 1981 (apparently, his first proposal was a Dreamlands supplement for RuneQuest, but they decided to go all in and just do a Cthulhu game whole hog. CoC also probably deserves a lot of the credit often given to Tracy Hickman for popularizing what later became known as the trad style of gaming in the earliest 80s. Don't get me wrong; it already existed prior to that, but because of CoC, Ravenloft, Dragonlance, and more, the style became one that publishers openly catered to and it basically became the mainstream style within the hobby from the earliest 80s to, arguably, still today.) Anyway, in the SPCM stuff, it's not really horror, it's really just straight up 5e D&D, which is pretty anti-horror, with Lovecraftian elements grafted on top of it, sometimes awkwardly, in my opinion.

And in fact that's what I want to get to. Sure, the Yig Snake Granddaddy campaign that I just read prominently features serpent-men, Yig (obviously), Elder Things and the Great Race of Yith. All of the SPCM Sagas prominently feature something from Lovecraft's Mythos; Ghoul Island was ghouls, Ghatanathoa, and Deep Ones, etc. The next one takes place on Yuggoth and prominently features mi-go. Etc. and etc. But the reality is that D&D always did have perfectly viable Lovecraftian entities that are unique, but just as good as anything Lovecraft himself wrote.

Mind flayers, for instance, first made their appearance in 1975 in The Strategic Review #1, the precursor magazine to Dragon Magazine. They became more official in 1976's publication of Eldritch Wizardry, the third supplement for D&D following Dave Arneson's Blackmoor and Gygax's own Greyhawk. Lots of other elements from the early game are vaguely Lovecraftian too; Juiblex and all of the oozes certainly had at least some nod to the shoggoths (also more immediately and obviously to the famous 1958 b-movie The Blob. The biggest problem wasn't that the elements weren't there, it was more that the assumed gameplay had little to do with Lovecraftian horror, and they were reconceived, as was everything in early D&D, as gameable challenges.

By the time we get around to thinking of doing something different, mindflayers, for instance, had been in D&D so long and were so familiar, as were other very strange "Lovecraft-like" monsters like the beholder, that they were familiar and didn't have that Lovecraftian feel to them anymore. Of course, the problem is that Lovecraft's own creations have taken on that same problem due to being heavily promoted in the gaming sphere. Call of Cthulhu is probably the most popular game (over time, at least) that isn't D&D or directly D&D derived), and Cthulhu himself, partly as a consequence of his saturation in nerd culture has become a bit of a joke or meme. The new minions movie in theaters right now for families heavily features Cthulhu, for instance. And a title that's clearly derivative of Dungeons & Dragons, for what it's worth.

So yeah, I don't think that Elder Things, the Great Race of Yith or mi-go are particularly compelling as Lovecraftian monsters honestly; if I'm going to do Lovecraftian, I'd just as soon see all new, all original monsters. And apparently, that's exactly what David "Zeb" Cook thought when he wrote "I1: Dwellers of the Forbidden City," one of his early products with TSR after being hired. This came out in 1981, and was specifically designed to be a Conan-esque jungle adventure mostly featuring yuan-ti, which it famously introduced. This is already Lovecraftian, or at least from the Lovecraft Circle, bearing more of a resemblance to Howard's creation from "The Shadow Kingdom" (alleged to be the first "proper" sword & sorcery story), but almost off-hand, as an aside, it introduced the aboleth. The aboleth was intentionally written as a nod to Lovecraft, but it's also a unique and new creature, not something like a D&D interpretation of a mi-go. Aboleth have kicked around in D&D ever since; they were added (like their co-module colleagues the yuan-ti) to AD&D explicitly in Monster Manual II, and have been in D&D ever since, getting pulled "forward" to the original Monster Manual for 3e and beyond. They also featured heavily in some 2e modules like Night Below, a big box-set module that's more like an entire campaign, where the aboleths are actually the end villains. They even featured in a novel trilogy during the 4e era. 

But clearly the best and most iconic aboleth product is the chapter on them in Lords of Madness, a late 3.5 era product that is considered by many to be one of the best products from the entire 3e era run (myself included.)

Now I'm not sure that aboleths have the same pop culture cache as some Lovecraftian entities, even within the D&D subculture. But they are just as good at being Lovecraftian as anything Lovecraft himself created, and actually, for my money, they're even better. They're not so alien as to come across as silly, but are sufficiently alien to come across as horrifying.