I don't think designers should be producing production-ready code at large(r) institutions.

There! I said it!

I think the goal of Figma Make/Claude Design/whatever Opus-4.7-driven-thing-to-kill-them-all you're using is to produce prototypes that look like your company is to do the following:

• Help with requirement handoff to engineers
• Explore North Star-type work
• Perform light-weight, not-in-app user testing before doing the laborious work of building new product.

I feel like there's this belief out there right now that these new tools mean designers should be shipping production code and it's the wrong use of everyone's skill sets. I think the packaging around how we want these tools to unlock designer workflows is the part that's confusing to everyone right now.

Most places I've worked, engineers are expected to own the code they author. This means that if something goes wrong, you're on the hook for fixing it. If you're a designer vibe-coding and shipping production code, who owns the code? Sonnet? (I'm being facetious.) (I mean AI is a better programmer than I am though.)

The burden of maintenance is ultimately going to fall to an engineer. So the right call is not expecting designers to vibe-code your production app, which dilutes engineering ownership in odd ways, but to use prototypes to understand the problem you're trying to solve.*

I think the odd designer in collaboration with their engineering team may occasionally ship vibe-coded work into production. I'm this kind of designer. It happens. But I don't think it should be the expectation, or the norm.

—Signed, a person who's been a “prototyper” at a bunch of places before AI made us all prototypers (the future will be pretty cool) (assuming the robots don't kill us all).

Should designers code 2026?

Happy vibe-coding!

* If you work at a 4 person startup, feel free to ignore everything I'm saying