Bing's AI image creator (powered by DALL-E)

Figured out the prompt for this position but you'd get a lot of dogs and sometimes it just doesn't fully unclothed naked people. oh wells. Sometimes it'l create hairy butts too.

View attachment 138857871
View attachment 138857831View attachment 138857841View attachment 138857851View attachment 138857861View attachment 138857881View attachment 138857891

Figured out the prompt for this position but you'd get a lot of dogs and sometimes it just doesn't fully unclothed naked people. oh wells. Sometimes it'l create hairy butts too.

View attachment 138857871
View attachment 138857831View attachment 138857841View attachment 138857851View attachment 138857861View attachment 138857881View attachment 138857891View attachment 138857901View attachment 138857911View attachment 138857921
what is the promts? original to create this image. on bing? please make available
 
It's hard, but sometimes I get some butt on Bing.

1722003911406.png


Prompt: German muscular man like bodybuilder, elderly, serious, huge muscles, huge legs, huge chest, veiny, ripped abs, handsome, huge, tall, dressed as bodybuilder. On a cabin on the forest. Bright day. Covered in olive oil. Back view. Wearing glasses and a tiny velvet apron. With his hands on his head. Flexing his back.
 
Chinese video AI Kling just dropped to the rest of the world. You get about 11 videos generations for free. Here's my experience so far.

Tried its Image-to-Video feature on some past gens, and overall it's better than Luma. Suffers from similar issues but better. Given what video AI was like only last year, the pace of development has been nuts lately.

Kling's image filter is fairly permissive. Roughly the same as Luma. Some pics I tried were allowed on one and not the other and vice versa. Seems to block pics based more on quantity of skin bared rather than individual body parts.

Kling's prompt filter, from my limited testing, is practically non-existent right now. Just avoid a few no-no keywords, and you can get basically anything through. Guess they didn't put much work much into the English prompt filter before they released it outside of China. However the dataset they used to train the video AI appears pretty sanitised, so I wouldn't expect explicit gens from explicit prompts using Text-to-Video. And when it comes to Image-to-Video, prompts don't appear to make a massive difference anyway.

Here's a few results. I cropped out the Kling watermark using ezgif.com in case they get shared outside LPSG. We don't need another journo writing up a piece about them, ruining everyone's fun, like with Cici.

ezgif-4-876372d360.gif



ezgif-4-5434d9b51a.gif



ezgif-4-2f05c53cf4.gif



ezgif-4-6623dd75d4.gif



ezgif-1-73ca93970a.gif


2.jpg


Optional prompt: Rugby player, prank camera footage, embarrassed, covering, posterior

1.png


Optional prompt: Rugby player shows off booty to tv camera

3.jpg


Optional prompt: Baseball player twerks his booty

4.jpg


Optional prompt: Camera catching nudie rugby player in shower, covering up his groin with a rugby ball

5.png


(cut gif short because rest of video progressed into eldritch horror)

Optional prompt: Rugby player lying on ground, while eager desperate crowd licking and rimming his booty
 
Chinese video AI Kling just dropped to the rest of the world. You get about 11 videos generations for free. Here's my experience so far.

Tried its Image-to-Video feature on some past gens, and overall it's better than Luma. Suffers from similar issues but better. Given what video AI was like only last year, the pace of development has been nuts lately.

Kling's image filter is fairly permissive. Roughly the same as Luma. Some pics I tried were allowed on one and not the other and vice versa. Seems to block pics based more on quantity of skin bared rather than individual body parts.

Kling's prompt filter, from my limited testing, is practically non-existent right now. Just avoid a few no-no keywords, and you can get basically anything through. Guess they didn't put much work much into the English prompt filter before they released it outside of China. However the dataset they used to train the video AI appears pretty sanitised, so I wouldn't expect explicit gens from explicit prompts using Text-to-Video. And when it comes to Image-to-Video, prompts don't appear to make a massive difference anyway.

Here's a few results. I cropped out the Kling watermark using ezgif.com in case they get shared outside LPSG. We don't need another journo writing up a piece about them, ruining everyone's fun, like with Cici.

View attachment 139176741


View attachment 139177101


View attachment 139177571


View attachment 139178141


View attachment 139178611

Thank you
(And holy crap, that last image!!! )
 
Gave Luma Dream Machine another shot after unsuccessful attempts with AI KLING. It seems the daily free credits weren’t available to me on Al Kling.

LumaDream.gif


This lovely AI still image was generated by another member on this thread. Sorry for not mentioning you but I lost track of your username. I used the still image in Luma to get the animation.
AI.jpeg
 
Gave Luma Dream Machine another shot after unsuccessful attempts with AI KLING. It seems the daily free credits weren’t available to me on Al Kling.

View attachment 139304551

This lovely AI still image was generated by another member on this thread. Sorry for not mentioning you but I lost track of your username. I used the still image in Luma to get the animation.
View attachment 139304581
Did you create an account?