Testing 4o Image Generation

Just wanted to show an example of the new capabilities just released by openai. It can now create an image and then generate subsequent related images. In the past if you tried to generate the same image twice it would always be different.

Lets build a plane

draw a design for a vehicle that flys and is like nothing we have ever seen. label the parts...

Can you now use the same design but show me what it would look i photographed in real life?

Lets add a pilot to this photograph.

😊 ...

The pilot is currently too big for the plane. He should not be outside of the cockpit

Meh, a little better. Must have some tiny legs or a really roomy cockpit.

Can we now create the engineering blueprints for this plane. Make it as detailed as you can following the original design.

So amazing.... My #1 tip is to take your time with the prompt. It really only does what you ask it to do so be verbose.