I know I haven't posted in a while and it looks like there's less activity here now. Is this sill an active place for AI wam?

It's the relevant place to post AI Wam. Whether it's active is more subjective. Creating good realistic AI wam images is somewhat trivial now. Personally, I just post something new of different now. I could post 100's of images a day but is there really any point to that. I don't see the point in flooding the site and creating storage issues.
My time has been limited recently and when I do, instead, I've been focusing on finishing my app. Tweaking settings, and adding some other functionality. Lighting and photography customisation, I'm also adding a clothing/outfit stylesheet generator.
One of the other things I've been working on is comic book and multi image sequencing generation. Using the api, give access to thought chains, image anchoring. This gives much greater consistency between images.

The samples are scaled down but each image generates in 4k and is stitched together automatically. I can generate 1-10 images or separate pages of 4-6 images each. It's currently pushing the limits of the model and there is some image drift which I'm slowly fixing with system prompts and anchors. If it gets good enough I'll release the app separately.

The reception here for AI was so negative that I've quit posting here much. I put together a compilation of some Val Day pictures then didn't bother. Its here now.

I think the negativaty has scared off others. When I do post I feel like I'm looking over my shoulder.

[edit add: I learned a lot here. And I check this group regularly.]

I don't even post finds in the "real" messy forum.

I think there are several reasons:

- Now the technology's been around for a few years, there isn't the novelty factor of seeing how close AI can get to being convincing.
- Since the generators are getting better at following prompts in reasonably straightforward language, there's less need to share the specific wording that generates a believable gunge tank or mud pit.
- Because of the above two, I think there's less enthusiasm for sharing imagery that might be seen as an inferior copy of what you can get for real. When this site's full of sample pictures of real models being gunged, a generated image of something that looks similar only not as realistic doesn't sound particularly enticing.
- It also becomes easier to generate images that are specific to the prompter's particular tastes, but may not appeal to a wider audience. For example, I've tried before to illustrate a mud/quicksand scene from the book 'She Died A Lady' (and attached another attempt here), but unless the viewer's read the book they probably wouldn't find it engaging. This also ties into the previous point - maybe AI-generated pictures of someone being gunged are of interest to me because she's dressed like Caroline Example, who was a hostess on Messy Gameshow X but always managed to stay out of the gunge, but to anyone who hasn't seen that show it's just another inferior synthetic gunging of a random person.
- There does seem to be a current of thought among some people that AI image generation is inherently wicked, for various reasons, but I haven't seen it show up in this group.

These days, if I post here the images tend to be scenarios which would be hard to do the old-fashioned way (such as historical or science-fictional settings, or requiring elaborate props) or where I've found the generator to exhibit some new-to-me ability (such as pastiche magazine covers / articles).

I think it's still active, less so maybe, but perhaps we're all just a bit more selective now.

You used to get posts with hundreds of near identical images where we were all just amazed by what AI was able to do.

These days it seems more about specific scenarios, or visual story telling.
Keeping the hundreds of images to ourselves & only sharing the 4 or 5 that get nearest to the image we were aiming for.

At least.. I hope it's still active, there's a lot of creative imagination going on here which we're unlikely to see done for real (without a Hollywood budget anyway)

@pie_okie, @messg, @sleazoid44, @uue404 @The Bootfull Dodger

I still visit the forums everyday and have to say that I am so grateful for all the content all of you and everyone else on UMD posts.

I've tried AI generation but don't get messy results at all so gave up. I love the images, gifs and videos you all create and am eternally grateful.

THANK YOU

Very interesting replies. I guess I'm a little behind the times.

Here's a better version of the lady in "uniform."

I think it's fair to say that the negativity before Xmas was around the usage of non consensual models and parties. Even excluding the ethical viewpoints, It's a fair stance to take given the growing realism and changing of laws in various countries around deepfakes. There is no general negativity towards AI as a whole in this group.
Outside of this group on the site, that's a different perspective and people are justified to having an opinion on AI.

There's a lot of negativity against AI around the whole site, but not so much in this group. I also haven't been posting much recently for several reasons. I've always been in a niche that did not get much attention here and I got the impression this also got worse over the last months.

Secondly I've been doing a lot of videos lately but it's a real pain to post them here. Too short to upload them as videos, too long to post them as GIF in reasonable quality without exceeding file size limits.

And last but not least, after more than 2 years of having fun with trying out several AI engines, watching them evolve and getting better, as well as constantly improving my own prompts, I feel I recently got a bit bored by it and even deleted everything I made in the very beginning. Still have lots of more recent creations lying around, but without anyone caring I don't really want to spend the time and effort to sort and upload them.

I see it similarly; I'm not uploading anything at the moment either. Besides, I tried everything I wanted to show last year, and it's become kind of uninteresting to me now. You know what works and what doesn't, especially since it's really just a hobby for me.

I still create dozens of similar images as I either tweak the prompt or edit the image to get closer to my (ever-changing) goal. I'll start with a specific outfit in a generic setting, or a generic outfit in a specific setting, then get more and more specific as I see things I either like or don't like in the created images, or have a completely different idea. But in the end, they'll all look very, very similar at a glance. I used to sort the images so I could post my favorites, often discarding all the ones that didn't look real enough or interesting enough to me. But now that realism as almost not a factor anymore, it's harder and harder to sort out the best, and like messg said, I don't want to flood the site either.

I still check in most days to see what others have posted, but... to be honest, I've mostly lost interest in other people's AI stuff. A lot of the people who post regularly seem to be posting variations of the same images quite a lot, having watched this group for the last two years or so.

On a mostly-unrelated note... Having spent so much time making my own AI stuff, I came to a really disappointing realization about real WAM: so much of it looks the same, just with different models. A few producers and models still stand out, but interestingly, a lot of them are lower-budget operations where the models or producers seem to just be doing what *they* want to see, not what will sell the best.

I guess that for me, that's probably the worst effect of having constant access to AI: it's jaded me. In trying to recreate my fantasies, I've become obsessively particular about the details, so much so that real-life shoots with minimal detail (nude, string bikinis, "cosplay" that is just bikinis with a prop, generic locations, outfits/costumes that are removed 10 seconds into the shoot/video, etc) have all started to look the same to me.

So... using AI a lot has actually led me to dislike most AI out there, especially in spaces where I feel it doesn't belong. Since I know how AI works, I can pretty easily recognize AI images and videos, and seeing a lot of people trying to pass them off as real genuinely bothers me.

As for the negativity surrounding AI on this site: most of the major concerns are legitimate, but both those and the people who hate AI haven't really affected this group much. The big anti-AI discussions have been held in other forums, not in this group, and have mostly been directly motivated by AI users breaking the site's rules, or by AI users posting AI in the non-AI forums.

Edited to add: I wrote this post when I was a bit down, and I hope no one took it personal. I'm just frustrated lately with real-life WAM producers or models I've adored going away from the things that originally made me like them. I understand why, though (appealing to the majority pays the bills), and I know that I'm part of the reason they're doing it. Not only have I stopped buying the sets and videos I like to help encourage them to make more like them (I can't afford them anymore), but I've turned to AI, one of their biggest competitors for engagement.

And on top of that, as Dungeonmasterone said, why look at other people's fantasies when I can make my own? I feel bad about it kinda detaching from this group, but I've just lost interest in other people's AI, and think this group has kinda gone away from "Look at these great ideas I've come up with and figured out how to create!" and instead is more focused on likes and views.

My preferences are a niche within a niche, which means there are probably less than 10 people (based on what I can tell from the likes) who actually consistently enjoy what I do. Many thanks to them for their support. However, as it's me who's paying for Google AI Pro (and also for basic access to Ideogram still), including buying extra credits, I almost entirely produce what I personally enjoy seeing and share things that I think are good enough to share.

My interest has switched from images to videos now, as I like giving some sense of "life" to the women in the images (if only for 8 seconds). Therefore, I now only generate images to use as establishing shots to make videos from. Whilst the clothes don't change that much, as I'm doing it to reflect my own preferences, I do tweak the scenarios to a certain extent as ideas come to me, but I'm not that fussed about striving to maximise variety. It's all just a bit of fun really and so I don't get too hung up about repetition and I don't feel like I'm spamming the forum, even though it's quieter than it used to be in terms of the number of posts. However, I sense that interest in what I do is gradually diminishing, so how much longevity my little niche still has is uncertain.

While it's nice to improve the realism of the images and videos, striving for perfection from that perspective isn't my goal really, not least because I don't have the time to keep doing what it takes to get to the next level when everything is improving and evolving so rapidly. Once others are producing wetlook images and videos of a quality that is consistently better than I'm capable of, then I'll probably gradually fade away into the ether!

Ultimately, I think to what extent people will continue to share their creations here is simply based on how motivated they feel to continue to make the effort to sift through and upload what they've produced, given that they're sharing it for free. It basically boils down to being a labour of love for them, so most content creators will probably be driven by what they think it's worthwhile for them to do based on the interest levels out there.

Omg I missed the latex ones. Love them! Wish they could pie them self in a gif

At one point, I was posting a TON of AI images here, but always as a theme or story. Then I did some video. The response was fairly lackluster so I eventually stopped posting at all.

Coincidentally, I just posted in the messy forum, an unedited photo of a 12 inch model and a short AI video of that same model in a realistic situation. Sadly, the hands are still the issue with AI, as there are glitches with the hands in the video. But I do like the overall result compared to a generated image. Again this was a photo converted to an AI video

I've been posting my AI creations on Instagram. I get more response (about 5 "likes" per posting, wet or pied) which is more than here and no "your putting real women out of work" crap. 177 followers.

I think the main thing with AI WAM is that it's invariably of more interest to the creator than anyone else, because it lets you enact your ultimate fantasy, tailored to your exact tastes, and as a result will only really be of interest to other who share exactly the same niche or sub-niche, and even then they can always create their own which will be of even more interest.

It's the same reason AI-generated wam videos will never replace real human ones - why buy someone else's fantasy generations when you can instead generate your own even better ones? Meanwhile real-world wam has the added selling point of "real experience", an actal breathing human experienced all the feelings and sensations of being pied or having their underwear filled with custard, which you don't get with generated stuff.

The novelty has worn off, and I think everyone's already seen it all before - I'm still sharing what I make and it has its fans, but it isn't getting interest in the same way that it was a year ago. Sora 2 has made mess clips trivial to make (even if a lot of them are rubbish), and images are old hat now. It feels like the tech has plateaued and the bubble will probably burst soon.

This is still an open forum for people sharing their creations though.

I half agree, the novelty of being to create anything is definitely not what it was but I also think there is still a huge amount advancement still to come both from Videos and to a lesser degree images.
I think Sora2 was a watershed moment. I don't think most people appreciate how far ahead of the curve the model was before they adde endless safety constraints on it. The same can be said of nanobanana pro for images. It is so far ahead of rival models it's silly but again, safety constraints pull it back.

The thing that's really struck me though is that there is a level of "good enough" most people reeach and are happy with seeing in Videos and Images. It fulfil's their requirement. Personally, I'm more interested in the prompting and application of the models now.

Nerd moment... As good as image and video models are, they're not genius' yet. They don't know what you are thinking.
So.. in order to help the model I've been building a prompt enhancer pipeline. Not just a singular system prompt but a workable Wam tailored pipeline.

I can provide the model a basic scenario outline and pass through a series of LLM renderings to produce a final YAML prompt.

User Prompt

MIG Builder (Takes the scenario and breaks down the core subjects, poses outfits, location etc)

MessInteractionGraph (2nd pass adds defined mess and interactions with subjects)

Physics Expansion

MaterialLibrary + GarmentLibrary + CoverageEngine

MessPhysicsGraph

Prompt Renderer

Final YAML Prompt

It's not there yet but does make producing good workable prompts so much easier and the output should work on most modern models.

I agree - and I think it's useful to remind ourselves that "real" photography is also just a representation of reality rather than reality itself. Gen AI has got to a point that we can get something in a "realistic enough" style that it works for its intended purpose, and for my preferences, I feel that more realism would start to detract from the effect.

I also agree that the way forward for what I do is to refine the prompting process and fine-tuning rather than wait for advancements in models that will probably be too expensive to run. I've taken to using a 'YAML style' approach to prompting (the actual syntax isn't YAML but it doesn't matter, it's just a way of splitting out the prompt into clear conceptual areas), and training concept loras so that fewer tokens can be used to get the desired effect - that is, baking the concept of "getting slimed with complete coverage" into the lora so that I don't have to waste tokens in the prompt describing every detail.

I'd be interested in seeing how your prompt builder script turns out.

Some interesting points here and to messg and thereald I've been following your work here for some time with interest.

I don't think AI can or will ever replace real human work and nor do I want it to, it is however great for playing with ideas. Look at it as an art or planning tool - not a replacement.

I've started using AI to generate images in recent times (mainly through ChatGPT) for various non-WAM stuff - I use it to help visualise ideas and designs, sometimes just to see how it might looks and also to help explain things to someone else. This might be a paint or styling job on a vehicle, a design concept or just editing an existing image to move stuff around.
They used to say a picture is worth a thousand words didn't they?

Back in the day I never learned how to use things like Photoshop and outside of grabbing a pencil the best I can do is crude stuff in MS Paint. For people like me this has opened a door that felt locked before.
Recently I did a design for a piece of clothing which started off the old skool way with a drawing then design outlines done in MS Paint. I fed everything into AI and it was quickly able to have a workable near-finished design plus model it on people in various scenes... Just don't let it think it can create a sewing pattern!
This has since been made into a real thing.

Once you know how a system works then yes you're going to have fun with it.
One WAM use I've found is to create storyboards for things I'd like to do. I've done it before the old fashioned way but this just makes it easier.

I'm not really interested in posting any of my work and doubt many would care; it's just nice to be able to let my ideas out now and breath. We all have a ton of ideas in our heads I'm sure.

That said I do appreciate seeing what some have done - either it chimes with my own tastes or sparks new ideas of my own.