You are viewing a single comment's thread from:
RE: 100 Steem Bounty: What is the solution to a lack of SMT development?
The biggest issue right now is that the reward pool is dysfunctional and a huge and increasing portion of it is going to nothing but SP holders voting for self-enrichment.
I couldn't agree more. The mechanism of incentivizing search for good content via the prediction game + curation rewards never really worked. Very small fraction of (arguably) good content gets rewarded, which in turn makes looking for it unprofitable. We at Steeve are trying to change that.
Here we described how. What do you think about this approach?
Posted using Steeve, an AI-powered Steem interface
I welcome new tools and initiatives but I don't think most of this issue can be addressed without changing the core incentivization model that is built into the blockchain.
Why is that? Where do you think is the problem with curation or what do you think would have to change?
Posted using Steeve, an AI-powered Steem interface
The incentives for self-voting and vote selling are too strong. Without addressing this, much of the inflation as 'wasted' and we end up with the perceptional and market disadvantages of high inflation without the benefits of a large treasury going to uses that actually help improve and grow Steem. Community initiatives (that I've seen so far) can only address this to a relatively small extent around the edges but don't get at the core issue.
You're right, self-voting incentives are too high and it's a big problem. On the other hand, curation never really worked even before the introduction of linear reward curve and consequent spread of self-voting.
So even fixing this wouldn't fix curation. For that, I argue, you need 1) better content discovery, so that people are able to actually find the good content even from unknown authors - this is a necessary condition for curation which has never been satisfied on Steem. Without it it's just circle jerking of known authors who have no competition and thus no incentive to continue writing good content 2) start rewarding the good content, not just good authors - this will incentivize both the creation and curation.
Posted using Steeve, an AI-powered Steem interface
Introducing linear isn't the only reason for self-voting and vote-selling. In fact incentives for high stake accounts or for methods of paying to concentrate votes were even higher under n^2. Making the payouts linear flattened the incentives but it really only did so by moving the too-high incentives from one group to another. They remain too high, but perhaps in a somewhat more fair manner (i.e. too high for everyone).
There have always been methods of finding valuable contributions, such as sincere efforts by large curation guilds, paid curators, and individual stakeholders interested in doing it. Maybe these would break down and/or need to evolve if the volume increased say another order of magnitude or more, but that has never been the issue we have seen in practice. It still does no good if the incentives don't support finding and sending rewards there once it is found, and they mostly don't.
The only method I know of that could possibly address the incentive issues (outside of a complete structural redesign) is strengthening and normalizing downvotes. You can't usefully downvote yourself and you can't really concentrate the rewards with them either. It is a bit like giving an executive a line-item veto but not the ability to spend money on his or her own. I can't say for sure it would work but it at least has a chance. There is a good deal of opposition though (in part because the 'normalizing' part requires a painful shift in cultural norms).
Unless we can try something (arguably) radical I don't really see curation having much of a future on Steem. Hopefully Steem can evolve into something else useful, like a neutral substrate blockchain for communities and tokens with their own methods of distribution.