As many of you are probably aware of, there has been a heated discussion all over Hive on whether the "Reward Pool" for Hive posts should be deprecated.
For context, the latest arguments for doing away with the reward pool are mostly downvotes and AI generated content.
Disclaimer
I'm writing this with the considerations of what was discussed in the first 2 hours of yesterday's town hall. It is possible that the below are just re-iterations of what was being talked about beyond that 2 hours of the meeting.
These are personal opinions expressed by me as a node operator that may turn out to be bad ideas to some reading this.
Arguments For
Removing the reward pool will reduce the amount of computation required for block processing. There will be no post payouts that have to be processed in every block and vote operations will be much simpler as it no longer needs to compute the potential payout amount, basically the same as voting a post past its 7 day window except that its rshares will still be used for rankings purposes (hot, trending etc).
This will speed up Hive node replay times post deprecation as processing these transactions are simpler. There will be fewer vops created due to lack of author and curation payouts (we probably do not need comment_payout_update_operation either due to lack of such payouts), which may slow down the growth of HAF databases given the same activities.
If post rewards are not reallocated to other categories (which I advocate for if this is implemented), it would theoretically reduce HIVE inflation by 65% (the current allocation), resulting in a projected inflation rate of 2.14% for 2024. The staking rewards for Hive Power is closer to 3%, making the base Hive Power APR higher than the inflation rate.
Unfortunately, this is where the advantages end.
Downvotes are still required
Unlike what most people who are in favour think, downvotes without a reward pool will still be required for moderation. After all, it is still a social media platform hence things like spam (though any means such as mass tagging), scams and phishing may continue to be rampant. These are the exact reasons why "report abuse" exists on centralized web2 platforms and downvotes need to exist in place of that functionality.
You can't beat AI
If you can't beat them, join them.
On one side of the spectrum, AI generated content could plague Hive if the abuse gets out of hand.
Meanwhile on the other side, AI could be used to write a killer Hive DApp or protocol and produce accompanying materials that can bring tremendous value to the Hive ecosystem. For instance, with the right tools someone could bootstrap a very addictive game built on Hive that works and scales well and produce very convincing marketing materials to accompany the product (the game in this case) that leads a huge amount of users (that create value!) into Hive.
Hence, the existence of AI does not justify removal of the rewards pool especially when it is used for good as described above.
Alternatives for content rewards
Ads on Hive
This idea is nothing new, it has been suggested by Khal in the town hall meeting.
As it currently stands, the current online advertising model does not work in the context of Hive. There is no proper decentralized and trustless way to calculate clicks and impressions accurately, and that's why there isn't a view count on Hive posts or a watch time metric on 3speak videos. Third party ad networks are a bad idea as this is one of the many ways people still fall for scams and phishing today, and the Hive community has very little control over what ad gets served. This is why people (including me) use adblockers and even the FBI recommends it for the same reason.
A potential solution is to build an ad network natively on Hive that we can control through upvotes and downvotes. Hivemind already has a get_discussions_by_promoted API, which shows that it does process transfers to @null to process post promotions.
What needs to be changed is instead of @null transfers, the HIVE could be sent to the content rewards pool (maybe deduct a burn fee that indeed goes to @null) and fund the content rewards pool that way instead of from HIVE inflation. And of course insert the promoted posts as an ad into Hivemind feeds.
The positive side is that the reward pool size is no longer limited to new HIVE being created. However, at its current state this is far from enough to replace the content reward inflation. You can see in the graph below that transfers to @null are next to nothing in comparison.
Layer 2 rewards
This is discussed often as the alternative to L1 rewards, however to me it is just shifting all the same problems associated with L1 content rewards into L2.
The only benefit is faster L1 block processing as described as the argument for removing the reward pool on L1 as the logic is offloaded to L2 nodes.
At its current state this is a terrible idea until Hive L2 infrastructure matures. Right now the only implementation we have is Hive Engine and it barely works due to various issues users have been facing lately. For now, the battle-tested L1 content rewards system works well as it is.
Subscription model
I generally dislike subscription models, but it could work in combination with Hive ads described above where paying subscribers (to a community/creator) can have an ad-free experience on top of perks provided by the community/creator. Maybe ad-free across all pages for manually funding the reward pool.
Can be implemented entirely in Hivemind by not returning any promoted posts when the observer parameter is a paying subscriber. The paying subscriber is identified by either a transfer_operation or fill_recurrent_transfer_operation with a specifically formatted memo.
Changing how we promote Hive
What is the very first thing that brought you into Hive? If your answer is content rewards or something similar, well that's the issue.
This group of people came to Hive mostly for the rewards, and will leave when the rewards are no longer there (either being downvoted, not receiving upvotes or reward pool being deprecated).
Take a look at some Hive websites, and think about what is wrong with it.
Earn money
Blockchain-based social network, join-us and earn various digital tokens, reward others and build a rewarding community.
- Front page of Ecency
REWARDS
By using the platform, users get rewarded in Hive tokens and can receive donations in our proprietary Speak token.
- About page of 3speak
Proof of Brain
“Proof of Brain” is a protocol that is behind the rewards of the publications of Hive’s social networks, where the authors of content receive rewards and the users who upvote those publications receive a part of that reward. A way to create cryptocurrencies with our creativity.
- Front page of whyhive.co, linked from hive.io
You have also probably seen places where Hive (as social media) being mentioned/featured outside the ecosystem having something similar wording where you can earn rewards for posting.
We need to change that. The value Hive (specifically social media aspect) provides beyond any rewards is a place for immutable content that cannot be censored once it is on-chain. It won't be easy to change the mindset of existing users here but it isn't an impossible task to achieve going forward.
Conclusion
Deprecating the rewards pool is more about scalability than tackling AI abuse or addressing the issue of downvotes.
If this is needed for Hive to scale, I'm all for it. After all, Hive is supposed to be Fast, Scalable and Powerful. Other than that, don't fix what isn't broken that is battle-tested over 8 years of this blockchain's history.