I had the idea for this photo for a few days when I decided to turn creating it into a competition. In one corner me with my camera gear, in the other one me with Midjourney.

From opening my camera bag to closing my laptop after a bit of color grading, the final photo took me about two hours to make. No advanced post, and as you can see I didn't even bother setting (or retouching) the hands to the usual 10:10.

In Midjourney I spent around 30 minutes with tuning the prompt and checking versions of versions until I had something I knew I can work with. Once I had the image, I spent around one and a half hours in Photoshop, turning it into this version that I quite like.

The hurdles and limitations of both approaches are quite clear and well documented by now, but I think taking a look at the time factor holds the key to what eventually ends up being an AI's job.

Little sidenote: I shouted at Midjourney a lot, and was quite bored by the ton of minute things I had to do in Photoshop. Meanwhile, as usual, I had a great time setting up the photo, playing with the placement of lights, creating the right shadows and reflections by moving things just by just centimeters.
Another photography versus A.I. experiment. This time no time limit and no edits on the outcomes.

Squint a little and the similarity is scary. But focus on the feel of the images, and the difference is quite fundamental. Strength, weaknesses and limitations of both production pipelines are well known, and there are no universal right or wrong answers.

But in the end the customers will only care about their gut feel. So no matter the technique, a firm grasp on how to direct look and feel is more important than ever.
Back to Top