This was successful but with the rise of more short-form content on the platform (content that can be read or viewed in less than 30 minutes), the community and the witnesses have come to a consensus that the 30-minute rule is taking curation rewards away from human voters who are actively consuming content and voting on material they like. For this reason, HF20 will reduce this window from 30 to 15 minutes.

I find it difficult to imagine that no one but me sees the obvious problem with this solution.

What human being spends 24 hours a day sitting on Steemit waiting for content? Does not the usual use case suggest that people have lives that they go about, and occasionally engage with one to two hour sessions – if that – of using the platform actively? Doesn't that suggest that a shorter curation reward window actually rewards bots far more than human beings, who are much more likely to discover the existence of the post or comment well outside of a 30 minute window. Possibly, given the amount of time it takes to recharge Voting Power, on the order of a couple of days between major sessions of engagement with the platform.

This rule seems specifically and explicitly designed to privilege bot monitoring of feed streams much more than human engagement. Even for short-form content, the discovery time is going to be well beyond 15 minutes.

I'm not even suggesting that this idea is wrongheaded. It's just wrong. It's exactly the wrong direction for solving the problem you say it's supposed to solve.

If you want to reduce the advantage that bot voting has over humans, reduce the advantages that bots have over humans. Make the curation reward window a day long, drop the decaying payoff aspect, and simply treat all upvotes within the window completely equally. Then humans and human curation actually has a chance to find this content and vote it up and aren't penalized for not being a bot with nothing else to do.

HF 20 is going to increase the number of bots that people use to get maximum curation awards. You get what you reward, and you are rewarding and automated process which merely looks for new updates on feeds that are likely to be popular, tries to wait for the optimum number of minutes, and then puts in a vote. It has nothing to do with the content. This effectively makes the content mean even less in the context of human interaction.

Congratulations. That's quite impressive.

This will better serve the original mission of the curation rewards budget: to ensure that the Steem blockchain distributes rewards to the most valuable content.

Wouldn't it be much simpler to simply remove self-voting from the system and maintain the rest of the architecture as is? That would immediately and completely reduce the author's rewards from any curational activity without pouring any funds back into the top of the hopper – in theory rewarding people who weren't even involved in either creation or curation of the piece in any way.

In fact, that would suggest that the greatest stakeholders have an even higher motivation to choke off rewards to people they don't like rather than reward content that they do like.

All of that money goes back into the hopper and has a much greater statistical likelihood of landing on their head, after all.

I reiterate, you get what you reward.

Stop rewarding self voting by making it impossible and you'll still get curation and creation. Simply "redistribute" the authors' cut that you think they don't deserve and you'll get less authors and more people interested in "redistributing" those funds largely back to themselves.

In hardfork 20, this “vote dust threshold” will be removed. After this change users with any amount of SP will be able to cast votes so long as they have sufficient bandwidth. Votes that are below the threshold will be posted to the blockchain but will have no impact on rewards.

Will accounts which actually possess SP be able to lodge effectively zero SP votes on content? This might actually provide us an opportunity to differentiate content which "more people should see" (social upvote) from "I think this author deserves some money for what they're doing."

And yes, those are different ideas and different things which cannot at this point be signaled to the system differently. For a social network, it's amazingly non-social.

Allowing that finer grained expression of intent will actually be a positive help.

... it is also important to disincentivize rewarding content with respect to which no other stakeholders see value.

Why?

This is a contention without any support.

Why would it be important to the platform to dis-incentivize voting up content that you legitimately see as valuable, no matter what other people think of its value? After all, the underlying assumption is that if I think it's valuable than somebody out there is also likely to think it's valuable and I should reward it.

The alternative is in many respects what we see now: chasing the Dragon. Everyone is looking to jump on board the next big viral hit and playing the numbers to do so because the actual content doesn't matter, rather than simply seeking out what they find to be good and valuable and rewarding it. Likewise, on the creator side, chasing the Dragon ends up with a vast number of crappy, minimal effort, minimal investment posts about cryptocoin and cheerleading for STEEM – which are rewarded because everyone sees that they are valuable in the sense of can easily acquire tons of upvotes, and you get the social network equivalent of incest.

That surely can't be exactly what you want. And yet – that's what you reward.

The changes required to add support for PoW mining for discounted accounts will be included in hardfork 20, but the actual PoW mining will be added later as a softfork on top of HF20.

Isn't this really just a backdoor way to get proof of work mining into the STEEM blockchain? With the removal of the powerdown restriction, even with the account creation fee getting burned rather than turning into SP for a new account, this just looks like one more way to get swarms of vote dust bots to leverage hovering around the SP baseline. Worse, the people that can afford to run bots in numbers are exactly the kind of people who can afford to run proof of work mining systems in order to create bot accounts in numbers at an even steeper discount.

Along with the changes to the payout timing window, this is one more place that more bot interaction is going to become ever more frequent rather than diminish.

I admit, I am not feeling particularly sanguine about these changes. From a user perspective, particularly one who is interested in the social network aspects of the system, these look to dis-incentivize interaction with the blockchain as a person, dis-incentivize being a creator even over what the situation is now, and open the door to even more bots at every turn.

It's as though the expected use case was never for people to be using the system at all.

I feel like that's kind of a problem. I may be the only one, but that's where I'm at'.