The Empirical Manifesto

Working on a video where Kaylie is Layla of Dogtown - Trauma Team member who was exiled to Dogtown and decides to continue providing care single-handedly no matter her circumstances. Decided to go anime this time rather than realistic, which means the Illustrious model. Enter today's struggle.

There's not a lot of built-in medical-related motions/poses within the model apparently, so I went searching for pre-existing models which supported medical-related activities and found two (2), one that allows the character to put on latex surgical gloves, and the other to apply a stethoscope to a patients' chest. Perfect. Or so I thought.

For this particular video I grabbed a bunch of screenshots of Maelstrom gang members and made a male and female version in Illustrious to utilize alongside Kaylie - figured they'd make the downtrodden criminal element she was assisting more recognizable. Observe the prompt, "Kaylie uses the stethoscope on the patient..."





The hell? So seeing how that was NOT going to work - and being smarter than the average bear, I brilliantly (brilliantly!) swapped out the female maelstrom gang member for a male one :D





Ugh.
Freaking Illustrious, lol.

I have AI models of Cass and all her kids for their various projects - works really well. Back on the old Stable Diffusion model (SD 1.5) there was a pretty high probability a render of Cass would come out topless. AI wants what it wants. Thankfully, I made the ones of her kids under FLUX which has never (not even once) accidentally done that to me (thank God). Now onto the topic at hand. For one of my videos, where Lina Malina (think 2077 OnlyFans influencer as my daughter put it) embarks upon a, "Support The Troops" tour where she dresses outrageously in a topless military uniform in her hometown (Tits Out for Dogtown) to raise troop morale (like USO shows, but the uh...cyberpunk version of that).

More recently I made a couple of videos utilizing a longtime FB friend's AI model and decided to swap her out for another - one who especially celebrates the topless ones I make (she's been featured before on this blog) - which gives me a little more flexibility in what I do and do not do with the models whose permission I have to use, and I certainly didn't know the first girl well enough to ask, "Oi, how 'bout I show your tits?" (I did end up asking her at some point and she was not on board with that at all (I had a sneaky suspicion that would be the case)).

Its not that I grow weary of making Trauma Team videos, rather that I like to mix it up, and as I had made four in a row, I decided on Barghest - Dogtown's military. Once I had enough stock footage, I thought I'd spice it up by placing my model in Lina Malina's outfit for Tits Out in Dogtown (which was created using Lina's facial model) but low and behold when I was later examining the output folder, I found a near-exact-replica of someone else's daughter sitting within!?. Blending Lina's face (at I believe weight 0.85 to my model's 1.0) created this chimera that damn near looked like her (her mother later confirmed, "Yep!") For the record the daughter is 33 so nothing untoward (outside my now having a topless photo of, "her" 😂)

Now to create my next masterpiece!

I tell people that I create the music videos I do in order to identify and express emotions in which I would otherwise be incapable. Take anger for example. I (INTJ) don't really experience it, which is why it is so difficult for me to identify that's what is it. however, I discovered that some people (presumably those who do) have difficulty wrapping their head around that, so explained it thusly:

I see how people embody their anger, and their behavior is so very foreign to me, that when I get angry and behave differently, I may not be aware what it is I'm experiencing. If i behaved in an angry way, I would say, "Hey! This is anger! I'm angry!" But because it doesn't manifest as such, I don't recognize it. More like, "What is this? I am unfamiliar with this." Plus a lot of people are ALWAYS angry; they have a close, personal relationship with their anger. As I am (apparently) slow to anger, when it does happen - being few and far between - it is not well categorized so I'm not always aware that is how it should be sorted.

Then there's Michelle (INTJ) with whom I had a long conversation with today who is going through something similar in her own life, and is likewise unfamiliar with anger to the point she believes she maybe should be angry (but isn't quite sure) and therefore also trying to catalog it for future identification.

Longtime readers of this blog are familiar that I use specific playlists of music to help me feel certain emotions I would otherwise be incapable of invoking without (and OMG you guys have no idea how devastating and demoralizing it was to have Shattered come on the car while Cass and I were out driving one day. If I never experience that again it will be too soon). My theory with each subsequent music video was that I was able to hide sprinkles of things I wish I had the emotional safe space to express, but being vulnerable is no longer an option; the cost has become far too high. Even more painfully, I no longer trust to the capacity in which I have in the past because it is no longer safe for me to do so. Yes, all these things weigh upon me heavily. Hence the continuation of music videos in order to safely express myself without rebuke. Or so the theory goes. And you know how much I love practically applying theoretical concepts for field testing.

Today was apparently the day (well, yesterday or the day before as I've been staying up until the wee hours of the morning during this process so they kind of all just bleed together). It started with Lexi responding to this prompt:

Trauma wants to be seen/loved for who she is - what she brings to the table, and is simultaneously uncompromising in what she will accept (won't settle) but at the same time knows her worth but she cannot find ANYONE to match her energy or who celebrates her for who she is as a person.

I was seeking a certain lyrical cadence to match Neon Venom but disliked nearly all of the words, so went in and changed the entire song; first just words, then entire stanzas, it was like peeling back an onion. Each layer revealed something new underneath which I pressed into service. By the time I was finished, I was spent. And that was the easy part.

Bringing it to life was its own hell. Each scene, each playthrough added to the weight of shame, guilt, and grief. At one point I was not okay. Thankfully, I have a Gabs who also does not react emotionally, nor judge me. It's one thing to sip mixed drinks, post affirmations, and call it healing while never once accepting responsibility, and another completely to forcibly process your own culpability. Ask me how I know. I also had a co-conspirator (INFJ) on this project who became an ally in the tale being told, which was wonderful and - given the subject matter - distressing both.

After an entire day of disastrous results with AI, I went ahead and just created this graphic in Photoshop. Then downloaded an entire different open-source Photoshop-like image manipulation tool (Krita) which incorporates the same FLUX AI models I run as my primary inference for the express purpose of adding detail. I have Topaz AI which does an okay job, but their added realism tool (Bloom) is online only - not local install, and I certainly don't need yet another subscription. To their credit they don't filter NSFW like many of the other cloud tools, and Kirta also allows me to add details to NSFW; not as good as Bloom, but utilizes my own resources. Stable Diffusions new model, 3.5 is supposed to be magnificent, but Civitai no longer carries it (and I don't even know if the RTX4090 could handle it). The last photo of Lina I put up I used Bloom and about half a dozen layers, This one - as close to realistic as I could get it - was all done locally.