If you want to pursue your own vision , it is better than you create your own fork under a new name.
I think his blog post there is pretty precise. Meteor has not moved much since then and the comunity is still fragmented (react vs blaze vs angular vs vue and ddb-pub-sub vs graphql vs grapher or even vs ddp-method-calls).
I think its now a chance for meteor to unify the community once again, go back to the roots, be more opinionated. Move good community packages to the core (grapher, a router?, redis-oplog (optional)), fix the datalayers problems (global nature of mini-mongo, no strong type support, sometimes awkward integration in ui-libraries) and have every important package be maintained.