The most dangerous thing AI does to our craft isn't replacement. It's the quiet conviction that the artifact on the screen is the same as the work behind it.
That is a striking article. I let myself disagree with you, partially. The confuse of form and fit was always there, it is proved by the tons of beautiful designs on Dribbble, it is proved by my own experience - every time I join the project, first I get - the folder with polished designs made by design agency, which will never be applied to the product because they won’t work. The AI makes this production process faster - yes. And that is actually good, because the beautiful but not functional picture now worth less. And good designs anyway need the final artefacts :)
The understanding that proper thinking and testing of wireframes makes not only good design, it makes the design that will work - that understanding was often a rear thing, and always a thing came via suffering from beautiful non-functional designs.
I believe there should be a rise of need in UX specialists sooner or later - to fix all the plausible interfaces :)
You’re right, and I want to say so clearly: form-without-fit didn’t arrive with AI. Dribbble has been a beautifully designed graveyard for a decade. The handoff folder of polished screens that will never ship is something every consultant on earth has received on day one of a project. The disease is old.
Where I’d push back gently is on the scale and the location of the damage.
Before AI, the form/fit confusion lived mostly at agencies and on portfolio sites. The cost was real but contained: a client paid for a deck, didn’t ship it, learned the lesson eventually. The bad work stayed in a folder. Now the same disease has moved inside the operating product. Generated UIs are being shipped, not filed. Generated PRDs are being executed against. Generated research is being decisioned on. The folder is gone; everything goes straight to production. That’s the shift I’m trying to name, and I think it’s a difference of kind, not just degree.
The other thing I’d add to your point about beautiful-but-non-functional being “worth less” now: I agree, and I think this is genuinely the good news. The market correction is happening in real time. The artifacts that were impressive in 2023 are commodities in 2026. What’s surfacing as actually valuable is exactly what you’re describing, the practitioner who can think a problem through, test it, make it work. UX as a discipline is going to be in remarkable demand for the next five years, and I don’t think the people predicting designer obsolescence have priced this in.
Concrete data point on your prediction, since you raised it: in the last six months we’ve added a new project type at the studio, basically we are finding ourselves fixing for clients the UX of Claude Code projects after the fact. Engineers ship a working app in two weekends. The flows are broken. The information architecture is whatever the model decided. The interaction patterns contradict themselves between screens. Nothing is technically wrong, but nothing is designed either. We come in afterward and do the thinking that was skipped. The meal is still made by hand, and there’s a whole new specialty emerging for the people who arrive after the fast-food rush and reopen the kitchen.
That is a striking article. I let myself disagree with you, partially. The confuse of form and fit was always there, it is proved by the tons of beautiful designs on Dribbble, it is proved by my own experience - every time I join the project, first I get - the folder with polished designs made by design agency, which will never be applied to the product because they won’t work. The AI makes this production process faster - yes. And that is actually good, because the beautiful but not functional picture now worth less. And good designs anyway need the final artefacts :)
The understanding that proper thinking and testing of wireframes makes not only good design, it makes the design that will work - that understanding was often a rear thing, and always a thing came via suffering from beautiful non-functional designs.
I believe there should be a rise of need in UX specialists sooner or later - to fix all the plausible interfaces :)
You’re right, and I want to say so clearly: form-without-fit didn’t arrive with AI. Dribbble has been a beautifully designed graveyard for a decade. The handoff folder of polished screens that will never ship is something every consultant on earth has received on day one of a project. The disease is old.
Where I’d push back gently is on the scale and the location of the damage.
Before AI, the form/fit confusion lived mostly at agencies and on portfolio sites. The cost was real but contained: a client paid for a deck, didn’t ship it, learned the lesson eventually. The bad work stayed in a folder. Now the same disease has moved inside the operating product. Generated UIs are being shipped, not filed. Generated PRDs are being executed against. Generated research is being decisioned on. The folder is gone; everything goes straight to production. That’s the shift I’m trying to name, and I think it’s a difference of kind, not just degree.
The other thing I’d add to your point about beautiful-but-non-functional being “worth less” now: I agree, and I think this is genuinely the good news. The market correction is happening in real time. The artifacts that were impressive in 2023 are commodities in 2026. What’s surfacing as actually valuable is exactly what you’re describing, the practitioner who can think a problem through, test it, make it work. UX as a discipline is going to be in remarkable demand for the next five years, and I don’t think the people predicting designer obsolescence have priced this in.
Concrete data point on your prediction, since you raised it: in the last six months we’ve added a new project type at the studio, basically we are finding ourselves fixing for clients the UX of Claude Code projects after the fact. Engineers ship a working app in two weekends. The flows are broken. The information architecture is whatever the model decided. The interaction patterns contradict themselves between screens. Nothing is technically wrong, but nothing is designed either. We come in afterward and do the thinking that was skipped. The meal is still made by hand, and there’s a whole new specialty emerging for the people who arrive after the fast-food rush and reopen the kitchen.