I think it's a worthwhile endevour. But the key is to look at how systems like the web stagnate and to design a system that can avoid the same fate.

Every project that grows large suffers the same inevitable descent into complexity. Things become so complex that it becomes hard to try out new ideas, which leads to stagnation. Eventually a better and leaner successor upsets the incumbent and the cycle restarts.

The key to creating great ecosystems is speed up that cycle. Design a system that encourages rapid growth and failure.

The 'web' should become a minimal hardware abstraction layer that offers safe access to graphics, audio, filesystem, and basic networking. Everything else could be built on top of that.

If someone has a new great idea for HTML or CSS just build it on top of the abstractions and hope others become interested. If someone wants to build a new browser they only have to implement the more manageable core.