You are viewing a single comment's thread from:

RE: The curation problem of HIVE. / Das Kurationsproblem HIVEs.

in #hive4 years ago

To be honest I was expecting more details in this post, but to respond shortly, yes. Automated voting compared to manual is way more profitable and manual curators are not only being disincentivized to read and curate manually but long form content is also being discouraged. People will either "manually" vote on something from a known author that looks long and decent without reading cause they don't want to miss out on the CR when they stumble upon those posts (I'm not completely innocent in this case either but I'm blaming my health reasons for not curating the way I have been before). So the 5min window is incentivizing short form content or content that takes little time to decide on the curation, such as art, a song, a poem, comic strip, etc.

You may understand that as a manual curator I really dislike autovotes, even though I see the merit they have the negatives that follow and affect most of the authors that get used to it and "abuse" it weigh down on everyone on the platform. This is why a lot of posts that trend have very little engagement and in a way it is not letting the hivemind decide what should be trending based on their interests and stuff they actually do read but who the big accounts, trails and front-runners are choosing at certain periods. At the same time the difference in rewards is very big, I recently went auto on an account just to test this and I was making over 4x more curation rewards than I am on my main account. It's backwards and will need some adjustments.

Another issue is the way curation rewards reward early voters and something that's been happening more and more often lately which is kind of backfiring instead of helping content discovery is being there early or being there early and it won't matter if someone else agrees with your curation to add their vote on top of yours cause the rewards you get are decent anyway. What I mean by this is that when certain big accounts see posts that already have $1-10 in votes they won't even care if the author cured cancer here and now, they are going to skip voting it to not risk being the last account to vote on the post which causes a chain reaction of other's doing the same. Often finding posts that have close to no rewards and voting on them to get their piece of the pie when many times there might be a reason that post is at 0 or close to it. Finding posts that are so low and exceptionally good is not only rare because there still are some manual curators of all sizes but as you say it requires work and people prefer to either be semi-auto or lazy. At the same time some accounts are choosing to curate these posts without spending time to actually curate, they just throw 10 votes onto whatever thumbnail looks good and as long as the post is under $1 or less and go on about their day - this is why using your downvotes will be more and more important for reward disagreement but not going to get into that now.

So there's more than one thing now disincentivizing long form content, both the 5min window and authors realizing that big accounts will avoid stacking votes on content that already has decent votes on it like the plague no matter how amazing the content is. This is also something I wrote about pre-HF and why I was a bit uncertain with your post to downvote anything making "over x amount of $" in the beginning of Hive if I remember correctly. I realize the point you had but to me it seemed like disincentivizing curators who don't care about maximizing their rewards and do vote late with downvotes and disagreement on rewards because they feel something is over-rewarded. This has now created a loop where the big accounts vote with tiny votes and spread them as much as possible, often also on posts that may not even deserve a small vote because they know they'll get the most out of it as curators. I've even seen certain accounts wait for an authors 3rd or 4th post to vote because they know many autovotes stop after 1-2 posts and they'll get the most out of it in CR with their identical 10% votes making so many posts hit the same identical rewards and along with no engagement and regard for the content you can see how automated, bottish it all looks even if it's manual. I don't have to name names, we can all go and check out how accounts are curating and when most of their votes are cast and how many other votes exist on the posts they vote.

Anyway, while this helps distribution it's pushing the autovoted posts onto trending because most of the possibly manual ones avoid voting late while the rest just focus on front-running or being the sole voters on certain posts.

Sort:  

It is an excellent answer/explanation @acidyo, and what I mostly want to add here is that it takes us back to the same issue that was always circling the drain on Steem: Is this seen as a content creation venue or just as a money dispenser?

Further to that... for the future growth and success of the entire Hive ecosystem, I think we need to be very careful in terms of how we describe Hive to potential new people. As I have said (about 6.5 million times!) as long as the primary sales pitch is "come to Hive to make money" the people who will be attracted to that will really not care very much about the content they create... they are purely going to focus on optimizing their rewards. And so, that particular snake will continue to eat its own tail.

So then the only way to change things is to start focusing on community and content creation. But there are a lot of people already embedded here who won't agree that's a good idea.

I have recently started creating posts that are entirely for "external" consumption, and it's interesting to see just how many page views some of them are getting. Now, whether that translates into new signups remains to be seen... but at least we have some tools now, like Hiveonboard, and such.

Yeah, I think it's pretty clear by now that we don't want to make the same mistakes as we did on steem in the first couple years, same thing regarding these conspiracy posts trending constantly. No offense and while we are an immutable platform we also focused on such and anarcho content too much in the beginning and too much of one thing is never good, especially not for a decentralized platform. So yes, while I agree that money shouldn't be the focus it is sad that so many here today and in charge of a lot of influence of the reward pool focus mainly on the ROI of stake instead of the value and outreach of it.

It's sad that those actively giving up on rewards because they want what is best for the platform have others riding on their backs indirectly taking their hive rewards and at the same time not helping.

What is your proposal to do next? Have the feeling the witnesses are acting the same as before on Steem. We had a little more alive HIVE at the beginning; Loved the first few days, maybe week or two or so when no auto voting system was connected to HIVE yet. I do think most poeple - in general - are good actors, but when the number of 'bad' actors are increasing, good actors become 'bad' actors as well. Thats what we've seen on Steem, and that is what we see on HIVE as well. I put 'bad' in quotes, since, what is bad? Imho, focus on money is bad for HIVE. Autovoting is bad for HIVE. For reasons as given in the posts and comments, like yours. I start wondering more and more if a social network with monetary rewards will ever became a success. A success in terms of money distribution to good content (long and short form) rather than to 'good' authors. One reason not mentioned why autovoting is bad: Any new user creating good content, not getting on autovoters, will leave as quickly as they came. Maybe they dont leave because they dont get rewards, but they for sure leave because the same users with good or bad content get 10 to 100 times more rewards. We are more than 4 yrs down to road and to me it feels like we hardly learned anything.

Loading...

Maybe one solution would be to change the curation calculation and remove the ordering of the votes entirely, so that it does not make a difference when you vote for content. Bots would not have a time advantage any more, on the other hand users would not be incentivized for being the first ones to discover new content. Or some mixed system, by reducing the importance of the ordering and also incentivizing late manual voting.

on the other hand users would not be incentivized for being the first ones to discover new content.

But maybe that would actually be a good thing. Why being first?
Right now a post that 2-4 days old is basically invisible and dismissed. Without the pressure to be first, the very same 2-4 days old post could actually still have a chance to be discovered.

I would love to see a random channel as the primary channel of posts. This gives more or less equal chance for a post to be discovered. Am being vocal about this various times in the past three years; Unfortunately none of the developers adopted this in their frontends.

You can sort the posts by time, which sort of gives you a random channel of posts: https://peakd.com/created

I know, but this is not the random we need. We need random across some time period. Maybe 7 days? This allows ‘older’ post to be discovered before they reached payout close time. Added to that, all blog interfaces sets Trending channel as primary channel. Back in the days SteemHunt experimented with the prime channel being random and loved it when they did that.

as long as the post is under $1 or less

This seems like the saddest part, from experience anyone posting regularly (and consistent quality) gets $1 within the first hour. However lack of comments can be more discouraging, especially for anyone used to other platforms. They might not even expect payouts, but the lack of visibility is kind of problematic.

Communities were a great addition, since they make it easier to discover interesting content. However, more personalized recommendations would be neat, other than the users we follow. Or maybe a way to filter posts, I remember curie has a website like that last year, filtering posts based on word count, pictures, words, payout, etc. Making that available to everyone would maybe help? People that are happy with their current earning they might not care about doing any extra steps, but smaller accounts might find content they can enjoy and interact with. Having more comments can be really motivating and finding like minded people.

Edit: I tried curating last year and it was so much harder than I imagined. Made me respect curators way more and it's really worth for others to give it a try just to see what it's like.

:/ Marky has me on an autovote that gets me $2-4 per post which has totally ruined my chances of attaining a vote any higher.

It's like he blacklisted me without blacklisting me.

I'm happy to bear that burden for you ;)