Steem has evolved dramatically from the initial implementation in 2016. However, one of the key features has persisted - stake-weighted voting.

The idea was based on the assumption that the highest staked holders of Steem (i.e. whales) would act in the best interest of the platform and delegate their stake to the best curators of the platform. Not just this - they were assigned exponentially more influence with an n^2 reward distribution curve. I.e. someone with 1 million SP will have allocate 1 million times more reward than the average person with 1000 SP. Not 1000 more as you'd expect.

Of course, such an assumption is utterly ludicrous and goes against human psychology 101. As expected, whales started acting selfishly in their best short-term self interests. In June 2017, the n^2 curve was changed to n^1, vastly democratizing voting. Of course, the 1 million SP still had 1000 times more influence. It still wasn't enough.

At the GOPAX meetup, Ned had a lot to say about this.

https://youtu.be/yxX_ttbXG-k?t=3851
(Skip to 1:04:11)

Firstly, Ned clarifies that Steem's algorithm was designed to reward the best quality content the most, and that is still the intention. So it's reassuring to see that he still cares about quality of content. Granted, that's a pretty vague goal, and not realistic, but there are magnitudes to how far one is from that goal.

Given they still have this goal, Ned was fairly contrite in acknowledging that the current algorithms have failed entirely. We can see this clearly - there's zero relation between quality and rewards. Indeed, over time, as Steemians have learned to exploit and abuse the system, there appears to be somewhat of an inverse relation at this time.

A few weeks ago, I made a comment that I don't see this system as sustainable, and it will need a radical solution people may not like. Ned seems to agree.

He identifies stake-weighted voting as the single culprit for this vast mismatch between quality and rewards. Instead, in SMTs, we'll have the option for a different type of reward distribution based on account-based voting. I.e. one vote means one vote, irrespective of how much SP one holds. There'll still be a superlinear curation rewards curve, so curators who find the post early will be rewarded much more. I'm assuming magnitude of curation rewards are still tied to stakeholding, but there's no mention of that. The first issue is - this totally kills demand for SP. But then, one can argue there are shitcoins worth far more than Steem that have no such concept similar to SP - it's all liquid.

Important caveat - this is an optional feature of SMT and may not be something Steem adopts. But let's assume it does.

The obvious effect is, it'll shut down voting bots, abusive whales, scammers, curation projects overnight. Each to his own. Of course, people can cooperate and collude, but there's no longer a way for one person to dictate thousands of times more rewards than everyone else. Of course, there's a massive downside to this - you are effectively empowering random spammers to allocate the same rewards as bonafide members of the community. Whales can make thousands of accounts to effectively Sybil attack the system too.

Ned does not address this issue, but I must assume Steemit Inc are working on a solution. My suggestion would be along the lines of active members from each community votes for curators, and the highly ranked curators then have higher influence over the reward pool. No such ranking system is beyond abuse, but there'll at least be some method to the madness. Furthermore, I'd also suggest a decentralized judiciary system to deal with cases of abuse. Yeah, I know Steemians will hate this. Of course, SMTs can take a more centralized approach and appoint curators themselves. This centralized approach will always be far more efficient than any decentralized method of curation. Yeah, Steemians will hate this implication even more. Sorry folks, but this is how human society works.

This is a complete revolution. There's a lot to digest here, so give it a watch, think about it, I'd love to hear your thoughts. Equally, there's precious little information, and it raises many questions. This was just a tease, I hope for a detailed post and whitepaper on the account-based voting concept from Steemit Inc soon.