Consider that, to do a new thing, we don't need a cohesive platform that replaces the whole of the web. We can leverage the web, e.g. use it to distribute native applications, or scripts that run on some new interpreter.

I like the look of the Red language. The goal there is to make a more modern, open source version of Rebol, which has aspects that tried to directly compete with the web some years ago.

I suspect the successful modern version of a Red-based open, distributed app network and hypertext document system will have more of a symbiotic relationship with the existing web.