I think we all agree that proof of stake is basically only a "solution" because we don't have any way to guarantee one person one vote reasonably. And even using KYC the way Voice does it will eventually fail because one person can create only one account buy can control arbitrarily many. Nothing stops me from paying a million people 10 minutes of their time to go through the KYC of Voice and give me their access.

I'm not sure that we can agree that is the primary motivation for Proof of Stake systems. I think it's safe to say that POS is largely in place for two reasons. Firstly, to encourage people to lock resources into the system and drive up the rarity of the underlying token so that its value is higher. Secondly, as an ongoing mechanism to draw outside investors in with the promise of being able to capitalize on that investment and take that value out along the way. The second one is actually the more dangerous of the two because we have simultaneously created an environment which demands investment come on in and a social order which is greatly against the idea of extracting value for that investment. The schizophrenia is generally dangerous.

That the blockchain only cares about pseudonymity and doesn't actually care about entity identity aside from the concept of consistency makes any kind of KYC system a bit of a joke, as you have already noted. On top of that, it's perfectly reasonable for you to pay a million people for 10 minutes of their time with your own resources to do something for you with their own resources. Hell, that's pretty much the definition of advertising.

The problem under the current architecture is philosophical rather than mechanical. POS coupled with Proof of Brain explicitly states that the people with more stake are/should be better at determining what is "good content" for everybody with stake in the blockchain. That's just an invalid assumption. It's just not true. Until you have a system in place that denies that to be true, nothing that derives from it is going to be useful. That's why I explicitly changed the terms from "determine the value of this post" to "give some of my resources to this post." The first is obsessed with minutia when it comes to deciding what an appropriate value is. The second is concerned with giving out what you already own.

If we take what you suggested, and I can just give myself 1 upvote a month with 100% of my own stake, and there are no downvotes to prevent this from happening would quite definitely mean the death of the inflation based reward mechanism of Steem.

If and only if we assume that the only reason that people interact with the Steem blockchain is to maximize the value that they can derive from it in a financial sense. If that is the only reason that people interact with the Steem blockchain, then absolutely, you are correct, the obviously degenerate play is to make one vote per cycle and to make it for yourself. All of your reward pool allocation will go back to yourself and thus maximize what you can extract for a given operation.

Making the degenerate play that obvious is actually an intentional side effect.

When it is that obvious, it becomes clear when there is nothing else of value to be derived from interaction with the platform. Though they don't generally have the vocabulary to say so, that is really the problem with the Steem blockchain as it stands. POS combined with a deeply complex bidding game which has no connection with the individual or content except as a means of trying to decide what other people are going to vote for means that the entire system is driven toward expressing a monoculture and deliberately trying to shrink the number of posts to increase the probability that whatever you vote for is likely to be what other people have voted for, unrelated to whether or not the content is good or you even care about it.

That is why I suggest getting away from that kind of architecture altogether and going for something that has a clear degenerate strategy.

What we need is a reason for normal, every day users to come and both create and consume content, not because they can make money (though that would be nice), but because they get an even greater value from interaction with the platform. That value can be the formation of communities and people of like mind, it can be the value of getting to see content they find interesting on a regular basis, it can be the opportunity to reach an audience – it can be any of those things and should probably be all of those things if you really want to have an underlying base of value for the token.

Because that is what we are missing. That is what the Steem blockchain has been missing for a very long time. Cryptocommodities in general have the same problem, so it shouldn't come as much of a surprise. There is no underlying thing of value for the token to be useful to be a signifier for. In theory, the Steem blockchain was supposed to be valued by the "value" of the content which lives on it. Unfortunately, because of the mechanisms chosen to determine what that "value" is, the systems pushed people into making very clear and sensible choices because the systems don't value individuals or content.

My personal theory is that it's because the original creators didn't have the imagination to consider that the decision-making architecture of the users will occur in an environment in which the system that Steemit Inc. designed exists. In a magical world where individuals interact with the Steem blockchain but their decisions aren't affected by it, the mechanisms in place might work. That world has nothing to do with our world.

Now, you are right, if we calculate every 7 day the split, that'd be an easy way to calculate the share nicely. But, the problem is that every vote I give, the interfaces will have to update the value of all posts I voted for already (bad overhead).

Except that's not true. Or at least it would be doing that kind of calculation anyway, except more of it, under the system you present. After all,, every time someone expresses an intentional value, the system has to recalculate whether or not your intended value has been reached, exceeded, or dropped beneath, and affect your vote intentionality accordingly.

Whereas under my suggestion, the system doesn't care about calculating that value for every vote you give – it only cares about calculating the value of the votes that you have given when a reward tick happens. Which is why I stated explicitly that you choose the length of your tick according to your computational resources. If the tick happens every hour, then once an hour you go through the list of votes, calculate the vote value (which is easy because it only happens on a per-user basis) distribute the reward pool, age out the oldest votes in the list based on whether they have gone beyond the voting window, and then do nothing until the next tick. A longer tick would require more memory overhead for storing the list of votes which would happen within that span, and necessarily affect more post entries in the database, but could functionally take longer because it's not as pressing. Choosing a one minute reward tick would decrease the number of votes that happened in its span and obviously decrease the number of database entries for posts that get touched, but the calculations remain very easy though the required time span for executing it becomes much smaller.

It's a scalable trade-off based on your available technology, expectation of activity, etc. On top of that, you can change it without any real repercussion as long as you do so on sequential blocks because it's an intermittent process.

While this is not that bad, I wouldn't like it personally, I think especially when you got a big stake to upvote, you might want to give different values to different posts.

Part of my objection to this is philosophical. If you have that large a stake, actively trying to micro game the system is a mechanical problem. It leads to an ever greater focus on automation, because there is no way that you can juggle that much and still have a hands-on, actual content-focused, interaction. I want to take that off the table altogether. If you have a large stake, that means that you can vote for a lot more things with a significant redirection of the inflation pool. Being able to direct how much of your voting power you allocate on top of directing where your votes go is just a recipe for what is generally thought of as corruption. It certainly leads toward a push for ever greater automation and a disengagement with content.

Again, the degenerate strategy becomes obvious. If you see a whale who is only voting for themselves, it's immediately obvious. You don't have to dig into how much is going where, you don't have to do any of that stuff. Degenerate strategies become obvious. (Strategies which involve mutual voting rings or pseudonymous multi agent self-reinforcement are an entirely different class of thing and much harder to ferret out. Trust me, I put in the time.)

But, for me this would definitely be something to have for the "smaller stakeholders" who don't care too much and just want to use the platform. Maybe a secondary system where you can choose "weight: dynamic" and one where you assign it specifically.

See, this is not how you design good systems. Now you've multiplied the number of interfaces and mechanisms. Now you have two potential mechanical sources of failure, and duplicated effort. That's not a good plan.

If you want people to engage with a system, you've got to make it easy enough to understand that the casual user will bother to, make the degenerate strategies clear and obvious enough that the big players will either not care and it's immediately obvious or expend the effort to come up with less efficient degenerate strategies, and you've got to make it easy to determine where the points of failure are.

Still, it doesn't fix the problem of:
a) Distributing Steem to as many people as possible to decentralize it
b) The problem of everyone just selfvoting
c) Grow this platform (Directly related to b)

You're assuming that these are problems. Perhaps worse, you are looking at the environment and coming to the same position that the original designers of the Steem blockchain came to and which is directly responsible for the situation being what it is today:

You believe that the most important thing about the value of the Steem blockchain is money.

That is not the case. No amount of distributing the token to more people will make it more desirable if the token is worthless. Self voting is not a problem (and honestly has never been a problem), but it's even less of a problem if the token is worthless. Growing the platform has nothing to do with self voting or with the way that votes happen, but it is surprisingly related to the token being worthless.

Elsewhere I have ruminated about the idea that the underlying value of the Steem blockchain is literally the "value" in the non-financial sense of the content that lives on it, and the reason that the value of the token is just above trash is that it accurately represents the value of the content on the blockchain as a whole. STEEM trades on the usefulness and value that people find in content available on the blockchain. If you really wanted to increase the value of the token, you have to increase the value of the content – not in a monetary sense but in a content-focused sense.

Your proposal explicitly doesn't talk about content. Economically, it talks about setting a price post hoc. You are still engaging with a betting pool, is just that you are explicitly stating what you think everybody else will set the value of a given post to. You have made the game exactly like it is now, just about rewarding players for an accurate guess orthogonal to any particular content.

The individual has no stake, ironically enough, in the content itself, and stating its value to them or lack thereof, it's quality or nature – it's just about playing with numbers. And that's it.

That does not solve the underlying problem the Steem blockchain. It increases mechanical complexity, it makes the UI even more ugly and nonsensical, and it doesn't provide an individual advantage to a common user on a day-to-day basis. It adds no value to their experience. It gives them no ownership of a resource, it requires more active management, and effectively it has all the failures of the current system, just with more things that can break.

I think the current system does a reasonable better job at A to C than the previous setup but still doesn't do a very good job at making people understand that downvotes are not mobbing but are a way to balance the payouts people get for the sake of the platform.

The problem for you in this particular paragraph is that, all too often, downvotes are mobbing. We have seen repeatedly evidenced that downvotes get used as a tool to mob people for personal reasons, and they are a great tool for that. They have almost no cost (and that cost has been decreased in the latest hard for work), they bang on an account's Reputation which is difficult to affect by the user, and as a mechanism for affecting payouts for the sake of the platform, downvotes just reallocate a tiny portion of the reward pool away from a given post to the rest of the reward pool – where it is almost guaranteed to that a spammer or exploiter has un-downvoted content which receives a portion of that down voted value anyway.

Don't feel bad, like a lot of people on the platform you have an unhealthy obsession with the financial side of things without actually understanding the process side of things. Things are as they are because people are involved, and the systems that have been suggested and definitely the systems that have been implemented, don't care about people. They don't care about individuals. They don't care about making people responsible for something of their own. They care about playing with numbers, and when that is the only game to play, that is the game that people will play on the platform.

Which is exactly how we got here.

Your proposal doesn't change that. It attacks none of the underlying architecture which leads to people engaging in what are largely negative behaviors. It just doubles down on some of them, in particular when it comes to downvotes.

That's not a good situation to be in.

Ultimately, I think everyone has to agree that the current mechanisms behind the Steem blockchain are untenable for a social media platform. At a very deep, basal level, they are directly at odds with the things necessary for a social media platform to grow and flourish. That problem can't be solved with the mechanisms in place because their basic assumptions are simply philosophically incompatible with getting the results we want.

It can't be patched to fix that because these are the axiomatic underpinnings of the whole system.

So it's going to go away.

The only thing that remains in question is whether or not one of the front ends will pull in enough standalone social media architecture systems to survive the process, putting a focus on content, putting a focus on user discovery, putting a focus on an individualized experience rather than the expectation that everyone knows better than you do – all of those things have to happen, and at that point really all you're doing is using the Steem blockchain is a storage database and maybe a way to keep score. Everything else will be done via some other database mechanism.

But it won't be the Steem blockchain as it's currently envisioned and as it's currently architected.