wow, one of the best comments I have read here since month. I almost lost hope that people here are actually into logic of distributed consensus.

I've been writing about this stuff, on and off, for a lot of years – some of them even before Steem was a gleam in anyone's eye. Which is, of course, part of the problem. A lot of the people involved in the current cryptocultism have no idea of the history of trust and trust-less systems, individuated versus top-down mechanisms, or the history of… Well, anything.

All of these ideas long predate what is essentially just a distributed database system with a point counter. That's a little bit of a problem for some people who don't have the context.

When I understand you right, your assumption is that there is either no collusion-resistant oracle OR sharding/super linear curves do not force consolidation of stake AND that we need proof-of-life to extract the action of the crowd instead of polling the "wise"-crowd.

My criticism actually goes deeper than that. Proof-of-life does no good when you're trying to actually determine what individuals want, whatever those individuals might be, whether it be a conscious human being, a tasked autonomous agent, or a collection containing one or more of either of the above. (And yes, I'm aware that I just referred to "collections" of agents as singular agents; sometimes when you are looking to provide what people are looking for, allowing agents to self-collect gives you leverage on being able to provide the ultimate goal.

We don't need proof of life. We don't need the action of the crowd at all. We need to ask the right question.

Predicition markets like futarchy are not perfect - true. They are practically collusion resistant. Proving individuals, --> needs a mechanism which averages out the user behind n accounts/sock puppets. You can average out by choosing some form of economic- or ecologic proof. When you go against economic proof, you remove stake... hmm now whats left? Attention? ID/bio-metrics? ...there are not that many smart proxies for uniqueness. One can analyse networking and connectivity - but this requires either semi-centralization/checkpoints/mods or a few users providing hard-proof/giving up anonymity.

This does bring the conflict at hand into a place where we can examine it.

Steem/Hive are not prediction markets. They pretend to be prediction markets of a certain sort. But consider what it is that is supposed to be "predicted"; the system is geared to try and predict the value of the content of a given post, with the underlying assumption that if everyone agrees on that value, that must be the value that it has, and those who have more stake have proven themselves to be a better judge of value and as such whatever they set as the value must be closer to reality.

You'll notice what is not brought up in that discussion at any point. "What I want" as a consumer and part of the system doesn't matter at all. "The content", whatever it happens to be, doesn't really come up or matter at all. Neither I nor the content matter to the process by which it will be ascribed valuation. What valuation is then returned, if it has no meaning in the context of being mine and no meaning in the context of being affected by what the context actually is?

The real application of the Steem/Hive model is as a betting pool on which posts the big players will find the most worthy of reflecting and collecting their big votes. To that end, all the content that wants to get the chance of being rewarded in any meaningful way is constructed purely and without hesitation to appeal to the ego or beliefs about the world of the set of major players. That's where the major stake plays and so that is what the majority of the rewarded community deals with. The betting pool only works as a betting pool if everyone involved has a covert knowledge of the kind of content which will be rewarded. I think we can safely look around at the last few years of rewarded content and say that things which are explicitly about Steem/Hive, generally rah rah boosting, and all about "I will teach you to make a lot of money for doing very little effort" is the kind of material that the whales have decided they're comfortable promoting at the top end. The travel and photo communities get by on the fact that there is an implicit statement that if you stay the course and play the game right, you, too, can be traveling to faraway lands and enjoying faraway meals.

A proper prediction market functions in such a way that the investment along the way represents a sacrifice, a trade, which can be cashed in when the single element that was predicted or not either happens or doesn't. If 65% of us are betting that Trump becomes president in 2020, that would generally be a pretty strong prediction. But if 65% of us are betting that I am going to like an article written at a fourth grade level which talks about how mean people are when he says how much he wants to hodl, you're going to be losing money. Because the results are about me and you could just ask me directly.

But then you would have to build up a new approach to understanding media, one that allows for the fact that I can have different ideas from other people in my community. That I can form communities which are not immediate but still share part of our existence and our tastes.

People are happy to tell you exactly what you want to know when it comes to "did you enjoy this piece of media"? They will tell you. Then you can remember that and use that knowledge to find other media that they might like, either recommended by a friend or having been sufficiently recommended by other people, have that piece of media brought up before them, and let them judge it in terms of value, keeping in mind the whole time that the specific value in play is going to differ based on situation, time, and who we trust with the knowledge.

Super short version: we need to stop pretending that we are deriving some sort of useful information from the global activities of actions on the blockchain and start acknowledging that nothing useful is possible without a re-focus on the actions and choices wants and desires of the singular individual. Give the individual the power to create their own webs of trust, to decide that they wish to signal to the system that this thing is better than that thing in their opinion, provide them easy abilities to make more content which satisfies them and pleasures others.

We don't need to prove uniqueness. We don't have to prove anything. We have to be able to observe a piece of content and then judge whether or not we believe that it would be good for a particular agent to be exposed to it and to decide on its value, keeping in mind the whole time that the value that gets assigned is just the value for that agent. Someone else would/could/probably ought to decide a different value.

We need to stop trying to decide if something is good just because enough people said it was good and instead shift to asking if it's good for X which can go either way.

We need to abandon the authoritarian, and until we do that. Until we get away from the idea that the wisdom of crowds should guide the lives of the individual, were going to be stuck here.