I think tooling (eg your verystable repo, or similar for rust/go/etc, or base library work like libsecp or the core wallet or libwally) and demos (your vault demo, ln-symmetry, spacechains, [etc?]) are better things to focus on. If the demos end up with us saying “this killer app seems ready now, and only needs this feature, not that one, and none of the other wannabe killer apps seem to have panned out yet”, then that’s fine. If we have working demos of useful things, that also seems like it would make it easier to say “oh, this alternative approach to implementing the feature makes things slightly easier/more efficient/more flexible” vs just having to go on aesthetics.
I think the advantage of APO/CTV/OP_VAULT vs the more amorphous collection of ideas is that it’s easy to decide/implement their functionality and then build applications on top of that functionality. If that’s true, then doing demos should be relatively easy; if it’s not true, then I’m not sure that paralysis in the meantime actually causes much problem.