AI Storytelling

Wow, more than a year since I last posted, aside from comments.
I've been doing some experimenting lately with using AI tools to write bits of stories. As I rather expected, the results are not something I would ever publish anywhere, at least not with my name attached, as I just don't feel that it's up to my standards. And the AI, no matter how carefully constructed my prompts, will insist on taking scenes in directions I would not have gone.

On the other hand - the AI does come up with an interesting idea from time to time, which may vary from my plans, but in a way which is occasionally worthwhile. And even when following my plan, it will occasionally come up with an idea or a twist which elicits a, "Huh. Yeah that's not a bad way to do it."

Of course, I do insist on giving the AI very detailed prompts. Because I generally cannot stand it when the AI goes off on its own into entirely unexpected territory, based solely on its extraction of patterns from other people's generic ideas of pornography. The few times I tried giving the AI some more freedom, I found the results to be quite disappointing.

So from my point of view, there does seem to be some worth to using an AI where I give it a fairly detailed outline of where I think the story should go next, and then let the AI give me its version of how that would work. The result won't be fully acceptable (and definitely not publishable), but it can provide additional plot points, paths, and twists which are worth considering for incorporation. Not that I'm actually expecting to write anything these days (my health is pretty poor), but I still like to plot out stories, and imagine the stories that I would write. And some of them do make good fantasy material.

More Thoughts on AI Art

A considerable amount of time has passed since I last voiced my opinions about erotic fan-art produced using AI tools. In that time, while looking over sites for erotic fan-art which appeals to me, I have necessarily been subjected to viewing a lot of such art produced with such AI tools. So, do I have anything more to add to what I've said previously?

Firstly, I still hold by the opinions which I have expressed before. The bulk of AI-generated erotic fan-art is boring and generally worthless for stimulating one's libido, or at least for stimulating my libido. Nonetheless, someone who is sufficiently imaginative is capable of using AI tools to produce erotic fan-art which is interesting, exciting, stimulating, or to which other positive adjectives may be applied. Such sadly doesn't happen often, but it does happen occasionally.

However, one new comment which I would like to add is that most such art suffers from the fact that the central female character is rendered sufficiently well, or even very well, but the male characters included in the image are virtually always generic rather than being taken from the male characters of the universe in question. This is something which is considerably more rare in art created by a human.

If I see a cartoon in which Akane Tendo and some guy are getting it on, then I like it when the guy is Ranma, or Ryoga, or Kuno (father or son), or Happosai, or her own father Soun, etc. I don't mind the occasional cartoon where Akane is sucking the cock of some random student at Furinkan High School, or getting gang banged by some unknown group of men who have abducted her. But I like it when the bulk of cartoons I am looking over featuring Akane (when she is not solo) also feature boys/men from the Ranma 1/2 universe.

Pink-haired Ranma

I was thinking of making some post wondering how long it would take for hentai images of female-Ranma to appear sporting pink hair, as she has in the reboot. But then I remembered that RanmaBooks had already made such a post back on Thursday, Oct 3.

Ranma 1/2 Reboot Starts

I saw the first episode of the Ranma 1/2 reboot tonight. Okay so far. I do feel cheated by one thing — we got to see female-Ranma nude, but not Akane. Or to be more specific, Akane was nude in a couple of scenes, but there was always stuff blocking the view of the most important bits. Perhaps the silliest part was when we saw female-Ranma nude. Apparently female-Ranma's breasts do not have nipples — we saw the entire curve of her breasts, but they were entirely featureless. I mean, if you're going to do that, then you might as well do the same as you did with Akane.

Other than that, no complaints about the single episode. I do wonder if they're going to follow the manga faithfully, or if they're going to include some of the filler which was in the original anime. And especially, whether they're going to be faithful to the characterizations in the manga. And most especially, whether they will show us the Akane from the manga, or the more violent and bad-tempered Akane of the original anime.

P.S. I'm hoping that this reboot will lead to a surge of new hentai art involving the Ranma 1/2 characters. Fingers crossed.

AI Rant

God, I wish we could ban all of the AI-generated garbage from our erotic cartoon sites! 99.99% of it is so bland and uninspiring that I can't imagine it producing an erection in a horny teenager. (Or the equivalent for teenagers of the female persuasion.) And it seems like every time someone inflicts more of this crap on us, they can't resist the urge to post a few dozen highly similar images, each one as worthless as the next. And the steaming piles of such so bury the remaining enticing posts by actual artists, that those few gems can slip by unnoticed as one's eyes skim through as quickly as possible in self defense of one's libido. I know it is theoretically possible to use sufficiently creative prompts to produce something with AI which can grab one's proper interest, but that accounts for the 0.01% leftover in the above figure, which to no degree makes me feel more forgiving toward the rest of the fecal matter.

AI Art

Expanding on a comment exchange from a previous post.

Do I like AI art, or do I disapprove of it? There are at least two things to take into consideration. (There are likely a lot more than two, but for the moment these two are the ones which occur to me.) The first is the quality of the art's appearance - how good does it look? The second is the content of the artwork itself - what is shown in the picture?

With regard to the quality of the appearance, I feel that a competent artist produces work which is superior to AI. This is largely because I feel that AI artwork looks rather bland. I suspect that this is because it is too uniform. A human artist will necessarily have more variation in the appearance of the art, some aspects will receive more attention than others, some aspects will be slightly less perfect. This can give the final image a greater richness of appearance. Although I grant you, a human artist of less skill will have an accumulation of such issues which falls short of what well-designed AI artwork can achieve.

However, I feel that the issue of overall quality is less important than the content of the artwork. This is the matter of the design implemented by the person behind the art, whether done by the artist himself or by someone directing an AI tool. What are the expressions and body language of the characters? What are they wearing? (If anything.) What are they doing? What is the setting? If there is any dialog, what are they saying? At least for erotica, this sort of thing can be much more important than how skillfully hands and faces are rendered. This is the sort of thing which captures the viewer's imagination, and either gets the viewer to immediately reach between his/her legs or just yawn and move on. When art design shows good imagination, the works of a rough artist who can barely sketch a human figure can possibly be more interesting or exciting than something rendered by a highly skilled artist. And in this regard, if an AI tool is skillfully used by someone with a good imagination who directs the tool in producing an interesting concept, then we the viewer can gain something worthwhile which we might not have otherwise been able to get if the person in question has little or no skill as an artist in their own right.

And so ... If we are talking about run of the mill art, nudes or standard sexual acts without much else going on, then I am generally opposed to AI art. In that case the quality of the art is the most important thing, and AI art simply cannot reach the same level of interest to the eye as a good artist. But if we're talking about art which captures and entices the viewer's imagination, then while a good artist is very welcome, I am also happy to receive the efforts of less skilled artists, or people who can get an AI tool to output what they have imagined.

Asstr May Be Gone

This summary is not available. Please click here to view the post.