Ever since generative AI blew up, people keep asking what designers should really be learning. My take? Don’t rush to master the latest tools—those things change every six months, you’ll never catch up. What you actually need is the ability to turn vague ideas into clear prompts, and the eye to tell good results from bad.
The first one decides if you can command the model efficiently, the second decides if you can pick the right image out of a pile of average ones and know what’s wrong with the rest. These two things are exactly what tools can’t replace.
As for specific skills—prompting, basic pipeline setup, a little bit of post-processing—that’s enough. Tools are consumables; judgment is compound interest. Don’t burn all your energy chasing the new stuff.
“Can’t keep up with the latest tools” — that’s way too real. Every six months there’s a new one, and by the time you learn the workflow, it’s already obsolete.
My approach is kinda backwards—I started by training myself to read images first. Once you can clearly explain what’s wrong with a picture, steering the model just clicks.