Generative expand blows my mind!
I was wondering what result I would get with a sharp foreground and a blurry background ...

Picture of Viktoriia Yatsentiuk on Unsplash

 Add a nice retro yacht

image (c) https://www.mollymccammon.com/

Greatly improved. From this:

To this:

Removed the obtuse tote and distracting bags below the table.

Expanded down first, then added legs to the barista.

Then expanded L/R.

Added some rocks to the scene, reflections are coming out nice!

Original:

Before Pic:

After Pic:

This would have taken me few minutes vs seconds with the new Generative AI

I needed a black glassball, but I did not have one so I "made" one.

A simple addition.

Yesterday, Gen fill stoped working as well as it had since it came out and was the beta version. I photograph events taken indoors on astroturf. Often, I use gen fill when I've not framed the subject and need to fill in the astroturf at the bottom of the image because I shot the image too low in the frame. Since yesterday, the AI isn't working as well. It now makes it look horrible. I've been filling the same exact subject matter forever. Now it comes out looking wrong. Sad that Content Aware fill now works better than the AI (gen fill.)  It's like they took my Christmas presents away from me.  Who stole my candy? This is all I do all day, every day with the exception of the handful of days a year that I photograph these events. So, it's not like I don't know the difference.  So sad that now 10% of the images I was able to use are no longer going to be viable and my workflow is going to take me so much longer again... I got spoiled!

They've done something to the "I" part of AI in the past two days. I've already been on a remote session with Adobe, so they've seen it.  The CSR did not have an answer and will forward this to the development team (I hope!)

I've had to roll back to previous versions of both.  Working fine again. Please tell me how to stop getting emails from this thread!!!