I've noticed a lot of "pie smash" A.I. images, and... something's off.

The pies are comically small, and -- more often than not -- if the pie is thrown, the plate hits the lady in the face, instead of the pie itself.

These last couple of days, I had the notion: what if they were already hit, having gone through Round 1, with plenty of pies in the background for Round 2, 3, and so forth?

I tried this prompt on Bing*, and it worked:

color photo by walter iooss jr of a 2010s tv scene featuring two beautiful 25yo blonde women in jeans, from the waist up; their hands are behind their backs. behind them are carts full of pies made of blue buttercream. There is an absurdly comical amount of blue buttercream all over their scalps and faces. no one else is in the shot. they give each other a mischievous look. white background low light film grain. the scene shows a fun mood

Obviously, if you want ginger babes, or Mexican cuties, or whichever, that's up to you. Also, the color of buttercream doesn't have to be blue. It can be green, or yellow, or pink, but I just went with blue.

I'll post the results later, but I wanted to let you know that there's a way around the "pie fight" A.I. stuff. You can give it a try.

*Yeah, I still use Bing! These last few weeks, it's gotten the job done for me.

that sounds very interesting

Sorry for the delay.

These were a few results that I had, including one where she's been in a pie fight, and is MORE that ready for round two

Just let your "mind's eye" fill in the blanks, as to how they got messy, and what may happen next.

I do have a bonus one, as you can see, where a lovely Mexican babe is surrounded by lots of "ammunition," just waiting to use it... or have someone use it on her (if you ask nicely)

That will be because Bing now allows you to choose GPT-4o as the image engine. This is the same underlying engine that Sora and GPT uses. The yellow tint on the images give it away. I don't know if the filters are different on Bing though. Either way, you might was well use GPT or Sora as you will have more control over the prompting and customisations.
The other thing to bear in mind is whilst you may get results with the current prompts you used, they are not natural language. you will get better results using more descriptive natural language