It would be fun to make a “mar-a-lago-meter” AI where you pass a photo, and it tells you its maralago score

I am thinking that:

• On the one hand, people complain about bloat. Do you need to have a full browser underneat? I mean, do you even need js support for most of the UI requirements be done with just html+js + passing callbacks to the underlying language?

• On the one hand, web frontend has proven to be so versatile. It makes moving applications from/to remote/local a bit easier, and you don’t have to learn “yet anonter script/language” just to draw buttons

I am not web or ui dev. But I think using html/css as frontend seems kind of reasonable, no? Expressive enough, simple enough, well known

What is the problem? Too heavy to interpret?