The Shape of things to come

[davegee]

Coming to Photoshop this year - available now in Photoshop Beta:

 

[davegee]

Another take:

 

[tom95521]

Generative AI is very interesting. The background fills are sometimes very good. Mona Lisa has been a favorite subject for fills.

Tested in photopea. One image I told it to fill with mountains. For the other image I did not tell it to fill with anything and it added some random objects. I have no idea what machine learning algorithm it uses to fill in with objects. The colors it chose are very good overall but the blending is not perfect.

AI is also being used to create music, design, and write. AI copyright is a hot topic.

Maybe some day PTE will have AI. "PTE, design a presentation lasting 3 minutes about my holiday in Hawaii using my photos and tropical music."

Tom

 

 

 

[rosy]

Hi Tom

There is a BBC Techy programme called Click, last week they had a very lifelike female robot creating a receipe. Don't know if this link will work for you ?

BBC Robot.

Rosy

[tom95521]

Hi Rosy,

Unfortunately my IP address is outside the UK so I would probably need to use a VPN server in the UK to watch the show.

Tom

[Barry Beckham]

 

[mhwarner]

I have watched many of the videos and I believe it will of great use to me in my travel photo edits.  I particularly liked the ability to move crowds of people from popular sites where I just didn't have the time to wait for everybody to clear out.  Or it will be useful for adding onto photos where I might have missed the end of a building or it got cut off because it was keystoned and I straightened it out.  My hope is that someday soon it will be able to do the Generative Fill from my own personal photos rather than pulling from Adobe Stock.  I certainly have enough to choose from.  I have seen that this ability will be available to enterprises to use their own branded stuff or images from a corporate owned catalog when the technology moves out of beta in the fall.  Hopefully we mere mortals will have some kind of ability to do this as well, avoiding any sort of copyright infringement issues.  I personally don't see much use to me for adding cats or dogs to couches.  Mostly I will use it for doing things that I can already do but which might ordinarily take me several hours of painstaking work.

Also, I did notice in one of the videos that when it comes out of beta, there will probably be some sort of charge for useage (credits or the like) in order for Adobe to be able to make payments to the artists / photographers whose images on Adobe Stock they are using.  Perhaps less expensive for private use than for commercial photographers.  I guess all that remains to be seen.

[davegee]

Generative expand sounds great for turning a 3x2 into a 16x9.

https://petapixel.com/2023/07/27/adobes-new-generative-expand-can-push-an-image-beyond-its-borders/

DG

[digartal]

I subscribe to Midjourney and the images generated are very photo-realistic. Midjourney now allows you to expand right, left, up, and down and then use PS to fill the 4 blank corners. Recently a lot of AI Upscalers have been developed and some images upsized to 150 cm for free and even bigger for a paid subscription.

Adobe Beta is also great at replacing content and reframing images See the examples below. The first one used the original image to change it with AI from 3:2 to a larger 1:1. THe second image used AI to replace my car in a 35-image 360-degree panorama.