You are viewing a single comment's thread from:

RE: Proposing A Worker Proposal System For Steem

in #blocktrades6 years ago (edited)

I think it should be possible to vote proposals that fund in either STEEM or SBD. For example:

> If stakeholders at some point decide they don’t like inflation being used to fund worker proposals, they can create and vote in a “burn” worker that pays a roughly equivalent amount of SBD back to the “null” account

In the circumstance when SBD is overvalued and in short supply it would be preferable to burn STEEM rather than SBD and in some cases vice-versa (although less important since the latter can already be addressed with conversions). There may or may not be other circumstances (apart from burn) when funding STEEM is acceptable or preferable. Possible examples: 1) advertising spend which can be easily ramped up and down in fiat terms, 2) a fund which can accumulates up to some desired balance, and the rate of accumulation isn't that important. 3) donations, prizes or other sorts of promotions where the fiat value doesn't matter much.

This proposal depends on SBD functioning well, so it should take care to help contribute to the successful functioning of SBD rather than potentially add further strains on it (e.g. burning SBD when it is overvalued).

EDIT: I have withdrawn the above comments after reading @yabamatt's reply which points out that proposal can simply be written which sells SBD then burns the STEEM (or can even do this conditionally based on the SBD price), much like my @burnpost project. Given that observation, the added complexity of selectable payouts isn't needed.

On the matter of inflation, I am totally against increasing inflation under almost any circumstances i.e. when not absolutely necessary. I do not believe it is absolutely necessary in this case. Such workers can easily be funded out of the existing reward pool, especially if they are only 1%. This amounts to reducing the reward pool from about 5% to 4% currently, which is hardly a catastrophic decrease (given how poorly the existing reward mechanism works, I'd be happy to reallocate even more to something that works better, but that's another discussion). It also could be feasible, I believe, for worker proposals to be voted that send funds back to the content reward pool, if that is what voters think is their best use (somewhat complicated perhaps by the fact that the existing pool doesn't have an allocated "system account" name like @null, although probably it wouldn't be all that hard to add one).

The reward pool has always (going back to the white paper) intended to reward contributions which add value to Steem, and the existing 7-day content-item voting method is only one way of doing so, not the only one (and not one that is even working all that well). Such worker proposals would amount to adding another method of awarding the same fund for the same (broader) purpose.

Reallocating rather than increasing inflation, apart from long term investor credibility issues, is as a practical matter by far preferable to adding even more supply and selling pressure into the market which has already struggled to maintain a non-declining price for Steem.

One last thing. I understand the historical context of Bitshares workers but I would strongly suggest to rename this to something like a funding or treasury or budget system. "Workers" is somewhat confusing and less compelling of a message to anyone who doesn't come from Bitshares, since the mechanism can fund actual workers but can also fund other things that aren't workers at all (including as you mentioned burn).

Sort:  

I would also prefer to reduce the reward pool to fund worker proposals rather than increase the overall inflation rate. I definitely like the idea of paying out in SBD rather than STEEM though for the reasons described in the post.

Regarding a situation where SBD is overvalued and in short supply, I assume in that case we can just not vote for the "burn" worker proposal, and vote in another proposal that instead will sell the SBD on the market for STEEM (assuming there aren't other proposals to do actual work that we want to fund).

Good point. The latter could be some worker version of @burnpost. I will edit my comment to note this.

Does reducing the reward pool include reducing payments to witnesses?

I would be in favor of reducing witness rewards proportionally with any other reward decreases, but keep in mind that the witness rewards are pretty small in the scheme of things.

Even if witness rewards were not decreased, I would likely donate funds on a regular basis to the pool and I'm guessing many of the other top witnesses and projects on the Steem platform would do the same.

I'm against reducing the witness rewards. They've already been massively reduced in HF16 and are making up only 10% of the total rewards in the pool.

However, with that said, I don't mind supporting some projects on a project to project basis (e.g. the 150+ STEEM bounty for the multi-sig implementation).

There's 2-3 witnesses in the top 20 that I haven't seen do anything meaningful in the last 1.5 years since I've been here. At least I haven't seen them do anything. I don't think it's fair to be taking the money away from stakeholders instead. By increasing the inflation rate, you are diluting the price of steem, which decreases the money of stakeholders.

The job of witnesses is to: a) faithfully sign blocks (not to sign blocks which would serve to impair the good functioning of the chain); and b) exercise good judgement in the interests of the blockchain and its stakeholders to vote to activate hard forks and set witness parameters (block size, APR, account fee, etc.). As far as I know all top 20 have been doing this. One could certainly question judgment in some cases, but that is pretty subjective.

Perhaps by "do anything meaningful" you are expecting witnesses to go beyond the above and provide some sort of bread and circuses, but that is not the job of witnesses, nor should it be. To the extent that witness elections devolve into a contest over who can do the most or spend the most, it destroys the security margin of the blockchain.

As a stakeholder I will vote against witnesses I see doing this. Likewise, I recommend that stakeholders focus their witness voting on evaluating the suitability and performance of witness candidates in performing the critical and irreplaceable functions described in the first paragraph of this comment.

I fully recognize that many stakeholders do not understand this and will vote otherwise, (most likely) unintentionally undermining blockchain. This makes me not all that optimistic about DPoS unfortunately, but wishing it were not so will not make it not so.

To bring the topic back to this post, I'm told that the very reason Bitshares implemented worker tasks was to ensure a clear division between witnesses (who do the essential witness tasks, see above) and workers (who "do" all sorts of things, as defined by worker proposals). (Bitshares breaks down the role even further into witnesses and committee but that's a discussion for another day.)

Interesting point with witness voting by stakeholders for me because how can someone MEASURE the performance in a very easy and transparent way? Because this is the basis for securing the chain. But if a stakeholder (regular user) can just do this based on gut feelings great marketeers will win which could be fatal long term in my point of view if the silent witnesses just do the crucial and most important things right (how to value for non-techie?) but seem lazy in the public view. It seems there is something wrong with perception of this task, how to solve this? How to value a good witness, reputation.. big no, could be fake due broken rep system, good posts, guess no because not the task, big stake.. big no because Intention not clear.. any useful key measures and numbers for that in a way everyone can easily evaluate to make the right decision without circle jerking and similar bads.

Posted using Partiko iOS

You are right. Big flaw in DPoS IMO. As you say marketers who have no idea what they are doing may win (and we seen that on Steem) and worse buying votes, cartels, etc. It seems that other DPoS chains have had more problems for whatever reason, but you are right to identify this as a concern.

With the grow of the chain witnesses should become "institutions" rather then "people".
Imagine a country where 19 people are gaining 10% of GDP, that doesn't looks right.
So later on I expect to see a witness to be a team of marketers, tech people and so on, later on it might even become something like some political party

double post deleted

In other words, witnesses should not speak out in favor or against worker proposals? Even if proposed a decrease in rewards for the witnesses?

Interesting, you should spread your opinion more often.

It sounds remotely plausible to me but I'm not entirely following your reasoning. Please elaborate.

Forgive me if I'm wayy out in left field with this, but isn't it fair to compare Witnesses to basically what the miners do for the POW chains, (but witnesses are for 'brain-chains'??)

It is similar yes.

what about producing blocks?
not meaningful enough?

Posted using Partiko Android

As a witness in the 90+ region I can tell you that I am burning money every month. My server cost me about 180 a month and I make 4-5 Steem a day.

Witness rewards are already likely on a trajectory to be much too small in Steem unfortunately. The way the system was set up, total inflation declines and witness rewards remain a fixed percentage of the total. So in the end state (18-ish years from now), witness rewards will be 0.09% which is a tiny security budget by the standards of most cryptocurrencies (excluding the ones with fixed supply but there continues to be more and more skepticism about that model working, for example see recent BIS Bitcoin paper).

For now, trimming off a bit of the roughly 0.85% witness rewards might be okay but as you said they're still pretty small in the scheme of things (and this is true both an absolute and percentage terms). I would not support a long term reduction in witness rewards though, as I think long term they are already on a path that is too low.

What about witness rewards being paid in SBD instead? The witnesses would prob like it under these conditions, but would be against it if steem ever went to $8+ again.

I don't think a fixed witness pay makes sense.

Part of witness pay (arguably the bulk of it) is for faithfully serving and securing the chain. If you have (hypothetically) $10 billion chain and witnesses being paid say $30K/year, that is a recipe for disaster not only in terms of scaling costs but also security. It may be fine when the value is only $100 million. But as the amounts of value at stake scales up, witnesses can easily be corrupted by outside influences worth more than their $30K/year position, but with significant negative costs for the success of the chain. You really want witnesses to be very well paid if the chain is worth a lot, ideally with more to lose from losing their witness position than to gain by acting against the interests of the chain/stakeholders.

I think a reasonable percentage per year makes sense, something likely in the 0.5% to 1% range, although the exact numbers may be up for debate in the future when we have more understanding of blockchains being used and critical to a rich, mature, economy.

i would also be interested (as said in my main reply) that post_rewards might automatically fund this worker fund :)

Maybe there would be a way to cohesively integrate witness rewards into the proposal and steem rewards economy. Witness rewards were always an incentive to for the infrastructure donation that witness have to make just to be witnesses on the blockchain. Maybe those rewards could serve more than one purpose that can be used to offset inflation.

It could possibly address the stale voting issue of witness votes by replacing the top X witness system with rewarding witnesses through funding appropriations instead. Without getting too specific, witnesses would be rewarded through successful projects that they fund through steemit appropriations. This would basically turn witnesses into angels. The more projects you lead to success, the better witness you are, and the more you will be rewarded.

I actually meant the opposite. Keep witness rewards out of the mix. As long as it is a small change I think most will be ok with it. I would suggest a 5-to-4 move is to big. Something like 5-to-4.75 would be a decrease in rewards by 5% and seems not drastic in either direction.

As I noted in a child reply here I would suggest against doing that (at least long term) but I would still prefer it to increasing inflation.

!ned

Enjoy your ned and don't forget to recommend nedcore!

up until now, if anyone says nedcore to me, I 'd expect to hear hard style music in few seconds...

For the reasons previously discussed in my post, I'd prefer to keep worker proposals paid in SBD, but it would only moderately increase the complexity of the system to support both. In such a case, the simplest implementation would be to have a separate pool of funds for SBD proposals and Steem proposals, with SBD proposals competing for SBD and Steem proposals competing for the Steem. A more complicated system would allow for some automated conversion between the two coins and all the proposals would compete with each other, regardless of the coin they were denominated in. In general, I prefer the simplest possible initial implementation without Steem proposals and leave it open to later improvements if it seems warranted.

I do agree with your concern about SBD price pumping causing potential problems, but I'd prefer to see implementation of the Steem->SBD conversion solution previously discussed as a method of stabilizing the price of SBD relative to USD.

I'm actually pretty neutral on where the source of the automated funding comes from. I suppose as written in appears that I'm favoring adding a new source of "inflation", but I'm actually completely comfortable with it coming in part, or in full, from one of the existing sources of inflation and I've been operating on the assumption that different people would propose reallocations from different existing sources as well.

On the naming issue, I would be fine changing the name. For the initial discussion, I felt it best to keep in the terms used by BitShares, as this would allow the greatest clarity for the reasonable size community of Steem users who are familiar with how it works and for people who want to take a look at how the BitShares system currently operates. But for long-term external marketing, I don't have any strong preferences for how it is described and probably some A/B testing should be used.

As @yabamatt noted, the SBD pumping issue can be handled outside of the core mechanism anyway. Proposals can be made to sell the SBD instead of burning it. This requires a bit more trust than funding directly to @null but that is true of the entire mechanism. So I have withdrawn that suggestion.

A more complicated system would allow for some automated conversion between the two coins and all the proposals would compete with each other, regardless of the coin they were denominated in

Although this seems to be off the table and not needed, I don't think it would be very hard to: a) keep the fund in STEEM, b) make the optional conversion from STEEM to SBD at time of payout based on a property of the worker. I.e. the same way content rewards currently work. The refund worker would be required to be STEEM-denominated in this model.

Yep, ultimately I think that would end up describing my second "slightly more complicated option".

I tend to fully agree with you on the reward pool adjustment over inflation increase. Well put.

Instead of increasing inflation what about reducing the interest slightly that is paid to the vested Steem daily? Wouldn't that keep the pool the same size and then have the funding come directly from the staked users who will benefit from the development?

⚡$$$ Tipped @tarazkp ⚡100.00000000 SWIFT! If you don't know how I work, reply help! Currently the price of SwiftCash in the market is $0.0019 USD per SWIFT. Current value of the tip is $0.1900 USD. To find out more about SwiftCash, please read our whitepaper!

That still effectively increases inflation since inflation that goes back to existing stakeholders isn't really inflation, it is more like a stock split. (In this case it is also done with the intent of incentivizing people to power up for various reasons, but economic effect once they do is as I described.)

couldn't some of that stock split be reallocated though instead without changing the available reward pool at all? I am just thinking that (I agree reducing the reward pool is a better option than increasing) reducing rewards might become a point of contention with smaller users whereas using a small percentage of the 'stock split' means that it will be funded from already staked users, meaning the largest will effectively pay the most even though on the same percentage cut.

Not my area though obviously.

The bottom line is that shifting from vesting rewards to anywhere else increase inflation on Steem Power holders. On what some other commenter noted is already a high inflation system, this is not a good idea.

I'm actually not a huge fan of the whole vesting system and I'd like to see it scaled back (for example, power down being 7 days instead of 13 weeks). If vesting rewards are reduced (or eliminated) then to avoid increasing effective inflation that should just go away (reducing nominal inflation) rather than reallocating it (increasing effective inflation).

None of the other current uses of inflation have this property so they would be fair game to reallocate, from the perspective of not increasing effective inflation. Which should be reallocated is obviously a matter of debate, even if we agree on not increasing effective inflation on an already-high inflation coin.

Or as @blocktrades originally proposed in the post, inflation (effective and/or nominal) could indeed be increased, but I'm personally very much against this for both long term and short term reasons.

Thanks for taking the time.

Has there been any talk of decreasing the powerdown duration of 13 weeks? It should be 2 weeks tops, or at the most 4 weeks.

It has been discussed several times but there is no broad agreement on it. My own view is one week.

I think 1 week is a little short, but 2 weeks seems good. 13 weeks is just crazy.