ChatGPT Images 2.0 is setting off the usual “design is dead” panic again, but the article argues the bigger shift is how quickly image generation is getting folded into everyday creative work.
Here’s the image from the article.
What worries me more than “design is dead” is the procurement/incentive shift: once a PM can generate 30 “good enough” comps in an afternoon, the org starts treating design like an infinite, low-cost input. Then the evaluation criteria quietly changes from “is this coherent and on-strategy? ” to “how fast can we produce options? ” and the person who used to protect the system (brand, accessibility, consistency, user trust) gets pulled into being a curator of machine output. That’s not the end of design, but it is a pretty big change in who gets to say “this is the direction” and what they’re rewarded for.
That “30 comps in an afternoon” thing is scary because it messes with the feedback loop: people start picking the shiniest render, then engineering has to retrofit it into a real layout system, and you end up with perf/accessibility debt as the hidden bill.
