Photoshop with AI Generation

Adobe has restrictions on what it will generate. Violence, some weapons it seems like guns, nudity, actual people and I am sure there are more. Living creatures like you and I, as well as animals the AI is horrible at generating
1689070388903.png


It can do somethings reasonably well, though I think the fact that it is not a real person or living being makes it horrifying in a way. There are always little things off with it as well. Below is "pretty face" completely generated by AI.
1689070507020.png


And just because AI can:
1689070727452.png
 
Consider what we were looking at 1 year ago, versus now...versus 2 years ago. The pace of refinement is insanely fast and I think within 2 years we will be a bit more humble in our criticisms.

What I said before (in another thread a few weeks ago) about AI/technology eliminating friction, so that people may more easily accomplish objectives. Eventually you'll notice that average people are producing higher end design jobs with less effort than it is taking currently. Whether that compute is being done locally or at a HPC center is I think the bigger question right now. Controlling the "production" of these physical items is much better than being a low or medium end graphic designer.
 
Consider what we were looking at 1 year ago, versus now...versus 2 years ago. The pace of refinement is insanely fast and I think within 2 years we will be a bit more humble in our criticisms.

What I said before (in another thread a few weeks ago) about AI/technology eliminating friction, so that people may more easily accomplish objectives. Eventually you'll notice that average people are producing higher end design jobs with less effort than it is taking currently. Whether that compute is being done locally or at a HPC center is I think the bigger question right now. Controlling the "production" of these physical items is much better than being a low or medium end graphic designer.
There is no doubt it will get better. AI generation last year was even more horrifying. Will this eventually kill the stock image market and kill off photography save for personal photos for personal events? Why pay for a stock image when I can simply have AI make me a royalty free image of 2 people shaking hands in a modern office
1689081508326.png


Or when you need that classic summertime shot for a beer ad, sitting around the lake enjoying a cool one
1689081817123.png


Now these pictures need work, but considering they are completely AI generated and in BETA, this will eventually get better and better and why would I pay when I can use completely original pieces. Now what would be something is if you could create a digital person and then have AI use that person in different settings, like having an actor for an ad campaign
 
There is no doubt it will get better. AI generation last year was even more horrifying. Will this eventually kill the stock image market and kill off photography save for personal photos for personal events? Why pay for a stock image when I can simply have AI make me a royalty free image of 2 people shaking hands in a modern office
View attachment 292118

Or when you need that classic summertime shot for a beer ad, sitting around the lake enjoying a cool one
View attachment 292119

Now these pictures need work, but considering they are completely AI generated and in BETA, this will eventually get better and better and why would I pay when I can use completely original pieces. Now what would be something is if you could create a digital person and then have AI use that person in different settings, like having an actor for an ad campaign
How much do you think the compute power is to generate one of these images?
 
They have to have some serious power behind them to do this on mass. I am sure there are tons of people playing with this every day and we are only in beta. Once this has wide release it will only go up from there. We are on the cusp of quantum computing so I would imagine this stuff would only go faster and faster.
 
In PShop the AI either happens on your computer or in the cloud - depending on what is being done. I use an M2 MacBook Air and it's quite zippy doing AI stuff.
You don't have to just create entire scenes. Sometime you just need to make some adjustments to an image.

For example, here is the source image ( https://tinyurl.com/yfuxruya ):

original shirt.jpg


The client wants him in a suit:

with suit.jpg


And a more vertical aspect:

Extended.jpg


With a different expression and eye direction:

Extended alt.jpg


Done by me (not a pro) on my laptop in just a few minutes. Is it perfect? Not yet - but this is still Beta. Even so, it's good enough that with some tweaking it's perfectly useable.

Don't forget the lesson of 1984:

LASERWRITER.jpg
 
Last edited:
Keep in mind that A.I. uses the web in some cases to aggregate a complete image of the information input into it. So, if you're generating an image, it may pull art from other derivative art or even copyrighted materials.
 

PressWise

A 30-day Fix for Managed Chaos

As any print professional knows, printing can be managed chaos. Software that solves multiple problems and provides measurable and monetizable value has a direct impact on the bottom-line.

“We reduced order entry costs by about 40%.” Significant savings in a shop that turns about 500 jobs a month.


Learn how…….

   
Back
Top