Generative Fill in Photoshop Beta

What happened with generative function?! I'm rare use it, but a year ago I experimented in many photos and it worked very well, but now incorrectly detects the light and focus. Maybe I'm not making prompts in the right way? But as I say before it was enough for me to write "flowers" and it understand how to draw.

Here I'm trying to add a vase with flowers on the table, it constantly gives me a sharp result.

First prompt was: vase with flowers. Second: vase with flowers on background, depth of field. Third: vase with flowers on background, not in focus, depth of field. And all the time same poor result...

For example, one of my experiment year ago with one of my photos. Here was generated: on table "book", "glass of wine" and "wine bottle", on background "dark transparent vase with dark flowers staying in shadow". All of this looking good! Maybe except glass, because no shadow from light but i made too small frame for it.

Why is Photoshop's Generative Fill (I'm on Photoshop Beta 25.7.0) so incredibly far behind the competition? Is it time for Adobe to give up entirely, or buy DALL-E? Here is my prompt:
"Photorealistic image of 12 people, aged 40 years old, doing olympic sports javelin, discus, high jump, pole vault, hurdles on the street Mount Pleasant in Norwich, UK while eating hamburgers and hot dogs and drinking wine, champagne and beer on a sunny day."

Using Microsoft Copilot (which uses DALL-E 3), these are the results:

Hilariously/disturbingly/horrifically, this is what Photoshop's Generative Fill, using an identical prompt (I just copied & pasted it) creates:

I think the results speak for themselves.

If I use Generative Expand leaving the text block blank, where does the image get its pixels from to fill the expanded canvas. Does it use pixels from the original image or are they generated by Ai

This should not of been released yet, in practice its extremely poor, it simply does not do what adobe claim, after about 5 differnt prompts this is still the kind of thing im getting to expand a background... basically just a stock image placed on a layer, this is supposed to exapnd the backround, after many attempts its not even matched one of the colours yet... 

I am frustrated and I think it is junk. I have tried so many terms to type to remove people on the beach and just leave my friend. Forget it. It just creates another person when you beg it to remove one. I say beg because I thought i was being clever when I tried every term i could think of. And when i go to report it, the window that pops up where you report it, it's too large and I cant get to the submit button. I try moving it up to no avail. I know AI is early, but Photoshop should not be advertising the awesome capabilites it does, for they are really not there. 

Remove is STILL a hit or a miss. So frustrating !

In August my prompts for white bunnies looked perfect and realistic. They would match the color tones and lighting of the image, now 2 months later they never fit the scene and most times look fake or cartoony. Why did the quality drop exponentionally in 2 months?

All that is nice.... if it actually worked.   I've yet to get anything usable out of this function.  Looks more like a childish cut and paste job.  And that's only on the results it actually follows through with.  90% of the responses I've gotten are error warnings of some kind of "community standards".  It's all but comical.   Apparently "wooden porch deck abandoned building" is against community standards. lol.

Don't blame the community for "standards" when the community didn't have anything to do with it.   Every time you say "community standards" as an excuse for results, we hear... "social credit score to limit speech".  But that's typical for Adobe.... you have to use a support "community" since there are no instructions for any programs you offer.  So you have to "earn badges" like children.   Kinda childish responses from the Generative Fill options.   Maybe if every word wasn't triggering to the "community" the AI would actually work.

I'm using Photoshop Beta 25.01 on Mac OS 13.5.1

When AI image is generated, a lot of distortion is occuring. For example, if I ask for a Bedouin tent in a Sahara desert setting, the tent edges are distored to match the sand drifts, as are any figures that are generated. In an image of a standing man, I generated an image of an office complex on a cloudy day — edges and face of man was distored (fragmented) as were elements of the generated background. I've tried this feature with numerous images and figures and it is always distorted. For figures I use the Select Subject from the floating bar, then inverse selection, then Generate. For landscapes I create a selection then go to Generate in the floating bar. I make sure my request are clear and no extraneous words are used. Too glitchy and honestly not ready for primetime.

Die AI generierten Spiegelungen im Wasser sind nicht korrekt