This makes a lot of sense: with AIs generating a large amount of code, the challenge moves towards validation and testing. If you generate towards a higher level of abstraction, there's less code and less moving parts to be reviewed and validated. Think of it as low-code/DSLs for GenAI, still code, but with less technical details to review and test.

Looks neat!As humans augmented with agents write more code, solutions that require less context shifting to get stuff done will win.A common web stack may include API handlers, OpenAPI spec, generated TypeScript definitions, generated TypeScript client, React logic and effects code, TSX code, HTML, and CSS.This generally needs filesystem watchers, code generators, transpilers, compilers to get stuff done.Something that can go from a backend handlers straight to terse markup for reactive UI would be a massive simplification to all this, and a big productivity boost.

it isn’t the concept that I find novel, at all. It’s the sheer depth and breadth of declarative expressiveness that I see in the examples. Maybe the best way I can put it is that what seems novel to me is how much novelty it appears to enable, before breaking through the abstraction to give it more capabilities.

I don’t think this is a reinvention of XHTML. It’s definitely closer in spirit to XUL, but seems different enough still that I wouldn’t call it reinvention.What seems particularly novel about this is that it’s taken the compositional approach of modern UI component libraries, and distilled it down to units of composition that can express a lot of behavior and logic declaratively. At least at a glance, that’s an impressive feat. Not necessarily in terms its technical capabilities (though that seems impressive too), but in terms of how it simplifies modeling interactive and data-driven UI.