What if we could compile designs just in time with generative AI? Given the current application state, the UI is conditionally rendered according to the output of some AI.
One more thought — generative interfaces might be best served as a “design in the small” paradigm. That is, instead of trying to generate an entire application, they might be more useful if they generate a single element or piece of the UI.
If the atomic unit being generated is a single UI element rather than an entire application, the monolithic software suite ceases to be the moat. Big tech’s form factor advantage erodes when generation happens at the component level. ecology-of-technology