But I'm not on par with that statement. SMTs and Oracles go hand in hand; and in the video - @ned and @theoretical sounded pretty serious about the math and logic behind account voting.
Ned really likes conceptual work. Which is great, I do too! But you can't really take them as signs that that thing is coming the way he talks about it. That's not necessarily inappropriate in this case, as SMTs are Hardfork 22 at the earliest, and that's probably next summer. (#21 is Hivemind.)
Anyway, thinking more about how this fits with my proposal, maybe I'm extrapolating what you're saying too much here, and please say so. But it sounds like what you would like is to keep the PoS function of Steem running but remove the content voting system from Steem entirely and migrate it to a new SMT.
I think that's interesting but it seems like it's essentially how EOS and ONO work and they're already there, or nearly so. Maybe that will turn out to be a better idea but I'm not sure getting into it a year+ behind is worthwhile.
I'm not convinced that stake-weighted voting fundamentally doesn't work, either. I think it's very appealing to have a vote that gains value based on how much effort you've invested here and how much you've contributed. That may be weighted farther toward external financial power than it would ideally be, and certainly the distribution of power is currently a mess, but I don't think those are insoluble problems.
My vision of Steem five years from now is one where there's a large population of established users with $1-$10 votes who can collectively guide the growth of the content and the community.