For a better Steem, remove the Steem reward pool

in #steem7 years ago (edited)

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.
Sort:  

I honestly think that you can only pick one:

  • Steemit with a reward pool or monetary rewards of any kind
  • Steemit where worthwhile content gets the attention it deserves

If there's something that Steemit has taught me it's the fact that capitalism and freedom fail. And I'm saying that after a decade plus as a capitalist libertarian.

I always thought that freedom would bring with it a beautiful meritocracy.

Instead, it's just an even harsher world of good brother networks, corruption, under the table deals, lying, scams, and several different varieties of bullshit only designed to benefit the absolute worst in society. I always thought that was just socialist propaganda. Turns out everything they have ever said has taken place on Steemit.

How tragic, now I need to find another ideology. This is legitimately hard for me.

I'm not giving up on the reward pool just yet. I'm still skeptical that a decentralized, monetized social network has any lasting appeal, but I'm not ready to dismiss it entirely having just seen one extremely flawed implementation of it. I still remember your post about your late girlfriend at the very top of Trending. It was definitely a case of "worthwhile content getting the attention it deserves". (Apologies if that recall was inappropriate.) I can think of many, many such examples from the past. There's potential, it just needs radical bug fixing.

About the failure of capitalism and freedom - I had written a post about this very subject several months ago. I used to be pretty open to differing political ideologies. I understood the benefits and pitfalls of capitalism, socialism; liberty, equality and everything in between. Unlike you, I didn't subscribe to an ideology, largely because they were all pretty flawed. However, Steem has been pretty definitive evidence that anarcho-capitalism is fresh hell. Good thing it's all a small, virtual experiment. (Actually, I take that back, I know there have been death threats and the like.)

Here's how I'll put it - if you believe in freedom, absolute libertarian models are at odds with freedom itself. That may sound like nonsense, but hear me out. You need to separate freedom and libertarianism. Absolute libertarian ideologies empowers the criminals, scumbags and lowlifes of the world. Think of it this way - they are encouraged to abuse others' freedoms. If you believe in freedom, I would suggest accepting good governance as the tool to protect people's freedom from the absolute worst in society, as you put it. That should be their only role. There's a bloody good reason governments exist in the civilized world - to keep the uncivilized in check. While governments are flawed - we all know that - it's sure as hell a lot better than criminals, despots or the underworld swooping in to fill the power vaccuum. Instead, what I would suggest is - don't give up on governments, but consider making better governments. The only solution to bad governance is good governance.

I have gotten a lot of hate in the past for suggesting the above, from Steemians, but I'll stand by it till people wake up and accept the reality. To be very clear, we are still talking about a free society, but one with checks in place to protect freedom and liberty.

Of course, if there were a species that's not comprised of selfish egomaniacs, and people were truly compassionate, I'd embrace absolute liberty in a heartbeat. Homo sapiens are not that species.

On a related note, I would totally like to see distributed judiciaries on Steem.

now I need to find another ideology

No you don't.

You are under no obligation to have an ideology. Admitting you don't know what would work to create a pleasant and balanced group or society is also an option.

Grasping what doesn't work is already several steps beyond what ideologists are capable of, as they are only concerned and stuck with a belief system they think should work.

That particular comment was tongue in cheek. And yes, well said. I just said somewhere that I've noticed myself moving more towards thr role of a general observer.

That particular comment was tongue in cheek.

I somehow suspected that, but I felt like crafting a nice comment anyway to say what was on my mind. Does that make you feel exploited? 8-)

Go ahead and exploit me.

.
.
.
.
.
.
.
.

That came out wrong somehow.

Be that as it may, I will still keep you to your word 8-).

Well, it's on the blockchain now.

The problem is that freedom is worthless without equality, and equality impossible without freedom. Take the "capitalist" out of "capitalist libertarian", and we're on the right track. At least that's what I believe ;)

schattenjaeger you totally right the little fish like me don't get any attention about my blog and i'm tryng hard...

I don't actually agree that the current rewards system has failed, but even so, there is always room for improvement. So it's not really worth debate on that point.

I have long wondered what would happen if post payout were modeled after something like a second price auction.

Generalized Second-Price Auction (wikipedia)

The generalized second-price auction (GSP) is a non-truthful auction mechanism for multiple items. Each bidder places a bid. The highest bidder gets the first slot, the second-highest, the second slot and so on, but the highest bidder pays the price bid by the second-highest bidder, the second-highest pays the price bid by the third-highest, and so on. First conceived as a natural extension of the Vickrey auction, it conserves some of the desirable properties of the Vickrey auction. It is used mainly in the context of keyword auctions, where sponsored search slots are sold on an auction basis. The first analyses of GSP are in the economics literature by Edelman, Ostrovsky, and Schwarz[1] and by Varian.[2] It is employed by Google's AdWords technology.

Second Price Auction (gametheory.net)

The theoretical nicety of second price auctions, first pointed out by William Vickrey, is that bidding one's true value is a dominant strategy.

I'm not sure if SMTs will let developers play with different payout schemes, but if so, I'm hoping someone will explore that avenue (and perhaps other possibilities arising out of auction theory that I'm unaware of).

I'm not sure game theory scales up to societal levels.

SMTs will definitely have account-based voting, which Ned has been hyping of late. I don't know the details, or the granularity available though. Either way, any SMT can build upon the code to code in their own customizations.

I'm not sure game theory scales up to societal levels.

Most of it doesn't; Evolutionary Game Theory might, but I don't think the original game-theoretical thinking that went into Steemit goes far beyond the one-shot prisoner's dilemma. Not saying that I would be able to come up with Something That Works based on Evolutionary Game Theory, but still.

@liberosist:

I'm not sure game theory scales up to societal levels.

Not sure how far it scales, but if auction theory applies for a massive use case like Google's AdWords (see the wikipedia excerpt in my earlier comment), I guess it should have something to say about a platform the size of Steemit.

I agree with this (@ocrdu):

I don't think the original game-theoretical thinking that went into Steemit goes far beyond the one-shot prisoner's dilemma

Although, to be fair, they did incorporate the idea of a reverse auction into the early voting penalty, so it eventually extended slightly beyond the basic prisoner's dilemma, and into auction theory.

Auction theory will most certainly scale up in size and numbers, but I doubt it has the necessary width and applicability when it comes to modelling group or societal behaviour that is more complex than an auction. What I mean is: I doubt it can model Steemit well enough to do predictions. I hope I'm wrong though, as it could help if it would.

I am glad to see you coming to this realization. I have been advocating for eliminating stake-weighting for quite a while. In off-chain discussion with @ned, he said he would address the issues I brought up before SMTs were released, and over the months since I became discouraged.

When he did address stake-weighting in his remarks in Korea, I was thrilled, and became eager to eat all my discouraged words.

"With no solutions in sight, the only thing that can be done is - cancel the reward pool."

I practically went apoplectic when I read this! =p Fortunately you went on to discuss various alternatives, and including what I have advocated, simply making upvotes not stake-weighted.

The rewards have proved to be paradigm changing regarding the quality of discourse and engagement on social media, and I see this paradigm shift as critical to the future. I would fain abandon it with the dirty stake-weighted bathwater.

Coupling some form of egalitarian voting, or rep-weighting, with elimination or dramatic reduction in bots and socks, would create organic valuations of content quality, and that is pretty much all I seek in terms of improvements to Steemit's valuations of content.

I have confidence that people are the best judge of what is valuable to them, and considering popularity to be judged by other metrics, such as stake, simply avoids people, society, being the judge.

For better or worse, people are what we are, and our judgement is thus what matters to us.

Thanks!

People have been questioning the stake-weighting right from the very beginning. I hope the noise becomes so deafening now that Steemit Inc and witnesses have to respond in short order.

I'm in the niche of weight loss and health and stumbled in here during the great flood of jerrybanfield in June 2017. At that time, he was pushing the idea of being rewarded for your content while gaining an audience due to steemit's incredible alexa ranking. I have received excellent rewards (comparitive to other places I post) and I have lots of ranking posts now, so that has been a win.

The cost has been a a huge investment of time and effort in "learning steemit." After about 10 years in this game, I can safely say this is the most complicated place I have ever posted - by a far, far margin. All the people I have brought in - about 20 content creators - gave up very quickly and think I'm crazy to be here.

I stay out of all the nonsense at the top end here and rarely follow anything about the future of steemit. But regardless of all the schemes going on. The fact is the we have less than 10K of "people who matter" on this platform.

The minnow kill rate of 95%+ is atrocious. I'm not sure if anything can fix this place when the masses are unable to participate. Every time I see a comparison to the early days of reddit and fb, I do not think it is valid. Those places welcome new users. We kill them when they enter through the door.

I watch this chart by @arcange and it is not a pretty sight. Most of the numbers decrease each time I look. One of them is growing quickly, and I'm sure you can guess which one it is.

arcange number of accounts May 7 2018.PNG

Are we looking at the same charts? When I compare an @arcange graphic to one from several weeks before it, yeah, the dead fish numbers have gone up. But so have the red fish, minnows, and dolphins.

That chart I posted is from May 7. This one is from 2 days ago, and yes, I do see some increase in some areas - nothing to brag about imo. But redfish (the one I really look at) was close to 170K in January and we only have vibrant growth in one category.

arcange number of accounts May 28 2018.PNG

To be fair, a vast majority of Facebook/Reddit/Twitter users are mostly inactive as well.

I am pretty sure they each have more than 10K in active users but maybe not. I'm following some people who are trying to work out global population figures and it turns out - most of us are non-existent.

You raise a lot of valid points but overall our system seems pretty decent and it does seem fully functional. If upvotes have value there will always be people who try and game the system. I think eventually things will work out and Steem will go to the moon! :)

I wouldn't bother spending a disturbing amount of time on Steem if I didn't see the potential.

Yeah me too. What do you mainly do on Steem? I tend to upvote and share my opinions a "bit". lol
You don't upvote yourself? Do you use dustsweeper?

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...

It may seem klunky, but the 13-week lock-in effectively reduces the supply of Steem available for trading and speculating. Increasing supply without simultaneously increased demand is a recipe for a price plunge.

I have discussed this both in the post, and below in a comment. On the other hand, it'll greatly increase liquidity, and the lack of a lock-in period will also significantly boost investor confidence. It used to be 102-week lock-in period before. The price was $0.10 when the transition to 13-week power downs happened. That's all I'll say, look at my response to dali.soh below.

well, i think that there are some vested interest in having a lock-in period for your steem in the form of SP. This is to prevent mass dumping of Steem on the open market that causes price volatility. Honestly, I would prefer to have liquid steem instead of SP to take advantage of any price fluctuation. Then again, having SP or liquid Steem have their own advantages and disadvantages.

Like I pointed out, the power down period changed from 104 weeks to 13 weeks at the end of 2016. It was meant to be a seismic shift, bring in mass dumps. Never happened. Half a year later, price was 20x higher than before that change was made. The point is - it locks in people from dumping when there's market panic, to be sure. But it also reduces liquidity dramatically, and the lock-in period has a negative psychological impact on crypto whales.

Also, pretty much every other crypto asset out there does not come with a lock-in period. Most are completely liquid. Yet, they are valued far higher than Steem, despite having a fraction of actual usage. I mean, many of them don't even have working products!

There could be a short term dump, yes, but I'm confident long term this will make STEEM a far more investor friendly token, and lead to orders of magnitude higher volume.

Your first point on the change of lock-in period of 104 weeks to 13 weeks which you claimed was 20x higher could be due to the rise in interest of crypto currencies. In fact, during 2017, the entire crypto market was on the rise, steem included.

I agree that having SP as a form of lock-in reduces liquidity. I totally agree on that. But then again, you could always choose not to power up and just keep liquidity steem. I think it's a fair for those who are investing in the steemit platform for the long term vs those who are just opportunistic.

I wonder if this is a mechanism created by dan as an experiment for EOS.

I cannot wait for the day wherein this platform will be filled with good content writers rather than a showoff place as to who gets trending. Your solutions in mind might work, hopefully. Or better ideas will be thought of @liberosist.

All your points are pretty valid but I think the only way to save the platform would be to heavily downvote and maybe even ban people who are abusing the platform but I am afraid it is not easy to do this on this platform!

Yeah, I would certainly like to see a decentralized judiciary system that lets you fairly ban people who are simply abusing the platform.

The REAL internet of content pays based on CPM (price per thousand views) and click-thrus ("Direct Response"). This is how Steemit should work. Make it equal pay per set of eyeballs or clickthrus, THEN, you can all discuss whether one set of eyeballs is worth more than another.

Curation happens automatically, it called click-linking, resteems, quotes, etc... This is how Google ranked pages in the very beginning, a Wisdom of Crowds solution. Adjustments can be made in the future to stop gaming, and those adjustments should be made periodically just as Google changes it's algos to reduce gaming of the system. It's never-ending, so witnesses will always have something on which to vote.

but you're not going to change ANY-thing, bc witnesses all have it too easy right now, and they don't want change bc it won't benefit them.
So they will vote all of this change down.

Steemit is sorta doomed in this way.

While not totally fool proof... what if something similar to consensus could happen? Instead of people having the power to give money to individual content, the blockchain does based on static reward amounts, and how the entire community feels about a piece of content.

A post is submitted to the blockchain.
Users can then vote for that post. Instead of having investment based voting which can be quickly purchased or controlled by those who are already crypto rich, the weight is determined based on your own blog contributions and the age of your account. However, this is a SLIGHT weight, not one where a single voter can totally over power the entire community if they choose. New accounts are now not hopelessly left behind with no way to catch up.

Votes are tallied for this post.
If the majority of the voters agree that this is indeed worthwhile content it can go to trending or hot based on some calculated algorithm involving the number of votes and the accounts.

Poster and voters are then paid a static reward based on these outcomes rather than the current vote weight system. Everyone who curates can receive an equal reward share, because the current system is confusing and discourages new accounts from participating. This gives everyone the ability to make meaningful votes, instead of just whoever has the most money.

If spam does get to the front page, it would be easier for the system to rectify it, and the majority of voters could decide that this content is NOT worthy of payment. If they do, it is removed from trending, and payment is nullified before the 7 day window happens.