As teams use AI tools to generate high-fidelity prototypes, the focus shifts to output instead of purpose.

Prototyping should be about testing ideas and learning what actually works before building.

annaecook.com/writing/2026...
https://annaecook.com/writing/2026/ai-prototyping-harder-worser-faster-wronger

AI Prototyping: Harder. Worser. Faster. Wronger. — Anna E. Cook

As teams use AI tools to generate high-fidelity prototypes, the focus shifts to output instead of purpose. Prototyping should be about testing ideas and learning what actually works before building.

Anna E. Cook
@annaecook my definition of "prototype" from a few decades in game development is largely the same: a prototype is something you build to answer a concrete, answerable question. if "can we produce polished looking art?" isn't one of them your prototype doesn't need polished art! definitely agree that various forms of high fidelity can just muddle the communication, the experiment, a prototype is trying to represent.
@jplebreton @annaecook
Lesson i learned in chemistry lab doing a “prototype a day” for 4 years is, “A prototype is an experiment to prove or disprove a hypothesis”.
In UX, it’s the same. Where’s the lab notes and conclusions? The experiment is a small part of that process.