After a few irreverent posts/comments and a shamelessly clickbait title right above, here's an entirely earnest one. It continues from a comment I made recently, in response to kevinwong's post and smooth's comment.
It's an universally held opinion that allocation of the Steem reward pool is faltering. Ned has recently restated the goals of Steem - i.e. a Trending page ordered by pending payout should roughly be in order of quality of a post.
[PS: To be very clear, by "Trending page", I just mean any list which is ordered by pending payout. Not necessarily the current Trending page on Steemit.com.]
That's pretty vague, and some of you may not agree this is a goal worth pursuing, but nevertheless we shall assume the goal of the Steem network is to allocate rewards by the value they bring to the consumer. On that note, Steem's reward pool allocation is an absolute failure - indeed, I won't be surprised if selecting posts at random (above a certain threshold of engagement) comes up with a better Trending page.
There have been several suggestions made, which may act as short term fixes. Unfortunately, many of them fall foul of the simple truth that hasn't changed since the advent of recorded human history - Too many people will always act in their own self-interests, over the interests of a society. This system will not be effective until every last major stakeholder acts altruistically. Mind you, these changes will help, but it's nowhere near enough.
Changing the author/curation split to 50/50. This will incentivize some people to do more curation than self-voting. But the rampant self-voters couldn't care less. They are getting 100% of the rewards anyway, irrespective of the split. Similarly, whilst bid bots will see lower bids, they will make up for much of the deficit with higher curation rewards.
Changing reward curves. This will help mitigate a lot of spam, but it'll also make the bid bots and self-voting whales that are currently dominating the landscape even more powerful. Something like an offset linear as initially suggested in early 2017 might be a better option. I.e. the dust accounts which tend to spam will have much of their influence stripped, but at the same time, the bid bots and abusive whales won't increase their dominance.
Investor class. This is a good one, but unfortunately, it is so exorbitantly profitable to delegate to bidbots and the like, that it will be a very steep cost (inflation) to incentivize whales adequately to keep them off the curation market.
Remove self-voting. This one's impossible, I'm afraid. They will just create a gazillion different sybil attacking accounts - there's no such thing as an "objective self-vote" on the blockchain. (I do have a solution in per-account voting power drain, but that's obviously going to be too expensive computationally for each curator-author pair.)
There are many more suggestions, but I think these are the big ones. Please point out some other popular ones I may be forgetting. While some of these solutions may improve things a bit, they also come with their caveats, and will ultimately not be anywhere near enough to effectively allocate the reward pool as intended.
We don't need to look for a boogeyman here. There's a very obvious reason why the Steem reward pool is so ineffective. It's a ludicrous idea that the richest people are the best curators (or have the best interests of the network) - that has been the downfall of many a societies throughout history. Most notably, of course, the Roman Republic.
Stake-weighted voting has failed. This is not a radical idea, people have been talking about this since 2016. Recently, Steemit Inc CEO Ned himself has publicly acknoweldged this. I don't really know what the popular consensus is - I assume stuff like this is discussed in private channels - but I know many who comment on my posts and at least one top witness agrees as well.
With no solutions in sight, the only thing that can be done is - cancel the reward pool. Gradually, over a period of time, downscale the Steem reward pool to zero. Stop printing SP, SBD (or STEEM in dire situations) to reward authors and curators, over time. Replace it with SMTs, which compete for the best method of allocating rewards. Needless to say, this is not a short term solution - it won't be worth considering till SMTs are ready for primetime.
Steemit.com can create their own SMT, which has its own set of rules, for a smoother transition. Or they can collaborate with Dtube, Busy etc.
Alongside trashing stake-weighted voting, Ned has been championing account-based voting as well. Note that account-based voting is a broad term which can lead to multiple implementations.
Let the SMTs compete for the best implementation. Some will think a completely centralized ecosystem where curators are chosen by moderators will succeed. Indeed, there's some evidence for this - the science community is perhaps the only one on Steem that effectively and accurately curates content according to the value they bring to consumers. They have achieved this through a heavily moderated curation program run by steemSTEM, which remains to this day the most successful curation project Steem has seen.
Of course, others would frown upon the idea of a centralized authority assigning curators and moderators / "oracles". And yes, banning miscreants, spammers and abusers from the reward pool. That's fine, they can create or join SMTs which have a more decentralized approach akin to what Steem is now.
Others may strike a balance. One idea could be a largely decentralized community, but with a voting system where people can vote for curators instead of witnesses. Something like a Delegated Proof of Brain. The top voted curators have the greatest influence on curating the platform. Of course, this needs to be backed by a strong reputation and engagement system, which decides the voters' influence, to minimize sybil attacks. At the same time, this remains largely decentralized. I'm well aware this can be abused, but it will still lead to a vast improvement, because the fundamental concept makes more sense than assigning influence solely by one's stakeholding.
Other SMTs may still choose to mix and match, have their own systems. The point is - who knows - through this process of free market innovation, maybe someone will actually come up with a solution that's sustainably and works for the many. All of this will tie right into Communities/Hivemind.
While we are at it, remove STEEM (I'll just call it Steem) Power and SBD. (Again, these are old ideas, people have been calling for simplification since 2016.) Make it all liquid STEEM. Investors hate lock-in periods and confusing multi-token networks. Consumers are confused, and to this day even some witnesses can't wrap their heads around the many asset types and how they work - let alone the average consumer. That brings us to the dinosaur in the room - will this not effectively destroy Steem's value?
The same was predicted when Steem Power went from a 104 week lock-in period to a 13 week lock. That was when Steem price was $0.10, people were prepared for a $0.02 price shortly. Didn't quite pan out that way. Yes, Steem's value might decrease, in the short term, as tokens are transferred to stronger hands. But we have to start thinking longer term - myopia has failed us too many a times.
Over the long term, Steem can be the core asset to many competing attention economies. There'll be ample demand for the strongest SMTs, benefiting Steem as a whole. Ethereum does not have any lock in, the core asset has zero value, and it's clearly worth dozens of billions of dollars despite very little real world use (lesser than even a flawed Steem network, certainly). Steem will become very much a core infrastructure asset like Ethereum is today. Indeed, many of the most successful tokens in recent times - EOS, ADA, NEO - are infrastructure assets with little inherent value themselves.
Of course, liquid Steem holders will vote for witnesses as on other DPoS networks; and the inflation will go towards interests and witness payments.
Finally, why not retain the Steem reward pool but transition it to account-based voting or some other form of reward allocation not as woeful as stake-weighted voting? That's a good one. Might even be implementable in the medium-term without as much disruption. Feel free to brainstorm.
Alright then, this has been an embarrassingly long rant, and I do apologize for the length. So here's a TL;DR -
- The Steem reward pool has failed not because of implementation issues, but due to fundamental showstopper bugs in the stake-weighted voting approach.
- Any fixes while keeping stake-weighted voting will likely only lead to minor improvements and will never solve the critical issues around human psychology.
- Transition STEEM to a core asset, let SMTs compete for a better solution.
- Remove the Steem reward pool, or replace it with a radically different approach of allocating rewards (curating content).
- Make STEEM a simpler, more investor friendly asset.