Though, due to positive feedback and winner-take-all effects in complex systems like human economies, I don't believe that's a stable equilibrium. The critical resources of systems will concentrate and consolidate over time. The question is whether there's any way to manage that in the architecture or protocol to minimize the resulting harm, or whether it's better not to try.
Once the winners take all, they tend to re-architect the status quo such that their power is easy to keep and hard to lose. Once they've done that, they don't have to worry so much about continuing to display whatever merits got them that power in the first place. No need to deliver on whatever promises made. No need to support whatever products sold. You're on top, your enemies are pre-crushed, you can now relax.
I agree that it's not a stable equilibrium. It'll take work to ensure that they can not in fact relax. But I think it's work worth doing. Much like how a farmer selects cultivars based on their desirable properties, so should should the masses wield their ability to revoke legitimacy and keep the position at the top precarious.
...which is why consistency is the wrong part of the CAP theorem to preserve. If your protocol cant represent inconsistency, then when power structures ossify and lose legitimacy, you're not left with a power vacuum. Instead you have the alternatives up and running and in use by a minority such that the only thing that has changed is that they're now the ones in the hot seat.
It makes peaceful revolution possible
Better, says my gut, to let either sides of a contradiction compete for legitimacy in the eyes of whatever local audiences are relevant.
Assuming consensus from the get go just doesn't seem to square with how large groups of people actually work.
So I hope that these alternatives work out for them because I'd like to have more examples to point at when I try to express this.